diff --git "a/wandb/run-20220202_223026-1j0459xm/files/output.log" "b/wandb/run-20220202_223026-1j0459xm/files/output.log"
deleted file mode 100644--- "a/wandb/run-20220202_223026-1j0459xm/files/output.log"
+++ /dev/null
@@ -1,14879 +0,0 @@
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-  1%|▎                                                   | 100/18250 [02:02<4:08:51,  1.22it/s]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-  1%|▌                                                   | 200/18250 [04:03<4:10:12,  1.20it/s]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-  2%|▊                                                   | 300/18250 [06:04<4:06:59,  1.21it/s]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-  2%|█▏                                                  | 399/18250 [08:07<5:07:58,  1.04s/it]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-  3%|█▍                                                  | 500/18250 [10:09<5:23:38,  1.09s/it]The following columns in the evaluation set  don't have a corresponding argument in `Wav2Vec2ForCTC.forward` and have been ignored: input_length. If input_length are not expected by `Wav2Vec2ForCTC.forward`,  you can safely ignore this message.
-***** Running Evaluation *****
-  Num examples = 4843
-  Batch size = 8
-{'loss': 3.3224, 'learning_rate': 1.8712499999999997e-05, 'epoch': 1.37}
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-Configuration saved in ./checkpoint-500/config.json
-{'eval_loss': 3.335383415222168, 'eval_wer': 1.0, 'eval_runtime': 191.0801, 'eval_samples_per_second': 25.345, 'eval_steps_per_second': 3.171, 'epoch': 1.37}
-Model weights saved in ./checkpoint-500/pytorch_model.bin
-Configuration saved in ./checkpoint-500/preprocessor_config.json
-Configuration saved in ./preprocessor_config.json
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-  3%|█▋                                                  | 598/18250 [15:32<5:14:11,  1.07s/it]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-  4%|█▉                                                  | 699/18250 [17:33<5:09:45,  1.06s/it]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-  4%|██▎                                                 | 799/18250 [19:36<6:38:28,  1.37s/it]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-  5%|██▌                                                 | 899/18250 [21:37<6:30:30,  1.35s/it]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-  5%|██▊                                                | 1000/18250 [23:38<6:23:33,  1.33s/it]The following columns in the evaluation set  don't have a corresponding argument in `Wav2Vec2ForCTC.forward` and have been ignored: input_length. If input_length are not expected by `Wav2Vec2ForCTC.forward`,  you can safely ignore this message.
-***** Running Evaluation *****
-  Num examples = 4843
-  Batch size = 8
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-Configuration saved in ./checkpoint-1000/config.json
-{'eval_loss': 2.93613600730896, 'eval_wer': 1.0000283848992335, 'eval_runtime': 191.7956, 'eval_samples_per_second': 25.251, 'eval_steps_per_second': 3.16, 'epoch': 2.74}
-Model weights saved in ./checkpoint-1000/pytorch_model.bin
-Configuration saved in ./checkpoint-1000/preprocessor_config.json
-Configuration saved in ./preprocessor_config.json
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-  6%|███                                                | 1100/18250 [29:49<8:02:27,  1.69s/it]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-  7%|███▎                                               | 1200/18250 [31:49<7:43:42,  1.63s/it]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-  7%|███▋                                               | 1299/18250 [33:47<7:44:57,  1.65s/it]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-  8%|███▉                                               | 1399/18250 [35:48<7:36:23,  1.63s/it]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-  8%|████▏                                              | 1498/18250 [37:50<4:51:04,  1.04s/it]
-  8%|████▏                                              | 1500/18250 [37:52<4:35:15,  1.01it/s]The following columns in the evaluation set  don't have a corresponding argument in `Wav2Vec2ForCTC.forward` and have been ignored: input_length. If input_length are not expected by `Wav2Vec2ForCTC.forward`,  you can safely ignore this message.
-***** Running Evaluation *****
-  Num examples = 4843
-  Batch size = 8
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-Configuration saved in ./checkpoint-1500/config.json
-{'eval_loss': 1.1156758069992065, 'eval_wer': 0.8358501277320466, 'eval_runtime': 192.2259, 'eval_samples_per_second': 25.194, 'eval_steps_per_second': 3.153, 'epoch': 4.11}
-Model weights saved in ./checkpoint-1500/pytorch_model.bin
-Configuration saved in ./checkpoint-1500/preprocessor_config.json
-Configuration saved in ./preprocessor_config.json
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-  9%|████▍                                              | 1599/18250 [44:08<4:42:57,  1.02s/it]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-  9%|████▊                                              | 1700/18250 [46:10<4:34:22,  1.01it/s]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- 10%|█████                                              | 1799/18250 [48:09<4:40:25,  1.02s/it]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- 10%|█████▎                                             | 1899/18250 [50:13<5:50:39,  1.29s/it]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- 11%|█████▌                                             | 2000/18250 [52:14<5:35:32,  1.24s/it]The following columns in the evaluation set  don't have a corresponding argument in `Wav2Vec2ForCTC.forward` and have been ignored: input_length. If input_length are not expected by `Wav2Vec2ForCTC.forward`,  you can safely ignore this message.
-***** Running Evaluation *****
-  Num examples = 4843
-  Batch size = 8
-  0%|▏                                                         | 2/606 [00:00<01:53,  5.31it/s]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-Configuration saved in ./checkpoint-2000/config.json
-{'eval_loss': 0.6002867221832275, 'eval_wer': 0.6314220834516038, 'eval_runtime': 191.4975, 'eval_samples_per_second': 25.29, 'eval_steps_per_second': 3.165, 'epoch': 5.48}
-Model weights saved in ./checkpoint-2000/pytorch_model.bin
-Configuration saved in ./checkpoint-2000/preprocessor_config.json
-Configuration saved in ./preprocessor_config.json
-02/02/2022 23:26:52 - WARNING - huggingface_hub.repository - Adding files tracked by Git LFS: ['wandb/run-20220202_223026-1j0459xm/run-1j0459xm.wandb']. This may take a bit of time if the files are large.
-Adding files tracked by Git LFS: ['wandb/run-20220202_223026-1j0459xm/run-1j0459xm.wandb']. This may take a bit of time if the files are large.
-Deleting older checkpoint [checkpoint-500] due to args.save_total_limit
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- 11%|█████▊                                             | 2098/18250 [58:29<5:47:32,  1.29s/it]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- 12%|█████▉                                           | 2199/18250 [1:00:30<7:03:49,  1.58s/it]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- 13%|██████▏                                          | 2299/18250 [1:02:30<6:50:16,  1.54s/it]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- 13%|██████▍                                          | 2400/18250 [1:04:33<6:41:06,  1.52s/it]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- 14%|██████▋                                          | 2500/18250 [1:06:33<6:35:33,  1.51s/it]The following columns in the evaluation set  don't have a corresponding argument in `Wav2Vec2ForCTC.forward` and have been ignored: input_length. If input_length are not expected by `Wav2Vec2ForCTC.forward`,  you can safely ignore this message.
-***** Running Evaluation *****
-  Num examples = 4843
-  Batch size = 8
-{'loss': 1.5812, 'learning_rate': 7.270153846153845e-05, 'epoch': 6.85}
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-100%|████████████████████████████████████████████████████████| 606/606 [03:09<00:00,  3.29it/s]
-
-Configuration saved in ./checkpoint-2500/config.json
-Model weights saved in ./checkpoint-2500/pytorch_model.bin
-Configuration saved in ./checkpoint-2500/preprocessor_config.json
-Configuration saved in ./preprocessor_config.json
-Deleting older checkpoint [checkpoint-1000] due to args.save_total_limit
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- 14%|██████▉                                          | 2600/18250 [1:12:41<3:54:28,  1.11it/s]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- 15%|███████▏                                         | 2700/18250 [1:14:42<3:51:06,  1.12it/s]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- 15%|███████▌                                         | 2800/18250 [1:16:42<3:55:25,  1.09it/s]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- 16%|███████▊                                         | 2898/18250 [1:18:40<4:04:12,  1.05it/s]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- 16%|████████                                         | 2999/18250 [1:20:45<4:47:43,  1.13s/it]
- 16%|████████                                         | 3000/18250 [1:20:46<4:42:22,  1.11s/it]The following columns in the evaluation set  don't have a corresponding argument in `Wav2Vec2ForCTC.forward` and have been ignored: input_length. If input_length are not expected by `Wav2Vec2ForCTC.forward`,  you can safely ignore this message.
-***** Running Evaluation *****
-  Num examples = 4843
-  Batch size = 8
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-100%|████████████████████████████████████████████████████████| 606/606 [03:08<00:00,  3.31it/s]
-Configuration saved in ./checkpoint-3000/config.json
-Model weights saved in ./checkpoint-3000/pytorch_model.bin
-Configuration saved in ./checkpoint-3000/preprocessor_config.json
-Configuration saved in ./preprocessor_config.json
-Deleting older checkpoint [checkpoint-1500] due to args.save_total_limit
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- 17%|████████▎                                        | 3101/18250 [1:27:06<4:40:17,  1.11s/it]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- 18%|████████▌                                        | 3199/18250 [1:29:04<4:48:12,  1.15s/it]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- 18%|████████▊                                        | 3299/18250 [1:31:05<5:52:11,  1.41s/it]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- 19%|█████████▏                                       | 3399/18250 [1:33:05<5:49:01,  1.41s/it]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- 19%|█████████▍                                       | 3500/18250 [1:35:06<5:43:05,  1.40s/it]The following columns in the evaluation set  don't have a corresponding argument in `Wav2Vec2ForCTC.forward` and have been ignored: input_length. If input_length are not expected by `Wav2Vec2ForCTC.forward`,  you can safely ignore this message.
-***** Running Evaluation *****
-  Num examples = 4843
-  Batch size = 8
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-Configuration saved in ./checkpoint-3500/config.json
-{'eval_loss': 0.40057530999183655, 'eval_wer': 0.38626170877093385, 'eval_runtime': 190.9848, 'eval_samples_per_second': 25.358, 'eval_steps_per_second': 3.173, 'epoch': 9.59}
-Model weights saved in ./checkpoint-3500/pytorch_model.bin
-Configuration saved in ./checkpoint-3500/preprocessor_config.json
-Configuration saved in ./preprocessor_config.json
-02/03/2022 00:09:41 - WARNING - huggingface_hub.repository - Several commits (2) will be pushed upstream.
-Several commits (2) will be pushed upstream.
-Deleting older checkpoint [checkpoint-2000] due to args.save_total_limit
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- 20%|█████████▋                                       | 3600/18250 [1:41:13<5:41:20,  1.40s/it]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- 20%|█████████▉                                       | 3700/18250 [1:43:08<3:06:07,  1.30it/s]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- 21%|██████████▏                                      | 3800/18250 [1:45:09<3:04:49,  1.30it/s]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- 21%|██████████▍                                      | 3899/18250 [1:47:09<3:11:38,  1.25it/s]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- 22%|██████████▋                                      | 3999/18250 [1:49:10<3:14:50,  1.22it/s]
- 22%|██████████▋                                      | 4000/18250 [1:49:10<3:06:08,  1.28it/s]The following columns in the evaluation set  don't have a corresponding argument in `Wav2Vec2ForCTC.forward` and have been ignored: input_length. If input_length are not expected by `Wav2Vec2ForCTC.forward`,  you can safely ignore this message.
-***** Running Evaluation *****
-  Num examples = 4843
-  Batch size = 8
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-100%|████████████████████████████████████████████████████████| 606/606 [03:08<00:00,  3.30it/s]
-Configuration saved in ./checkpoint-4000/config.json
-Model weights saved in ./checkpoint-4000/pytorch_model.bin
-Configuration saved in ./checkpoint-4000/preprocessor_config.json
-Configuration saved in ./preprocessor_config.json
-02/03/2022 00:23:55 - WARNING - huggingface_hub.repository - Several commits (3) will be pushed upstream.
-Several commits (3) will be pushed upstream.
-Deleting older checkpoint [checkpoint-2500] due to args.save_total_limit
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- 22%|███████████                                      | 4099/18250 [1:55:31<4:11:12,  1.07s/it]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- 23%|███████████▎                                     | 4200/18250 [1:57:33<4:03:11,  1.04s/it]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- 24%|███████████▌                                     | 4298/18250 [1:59:31<4:12:13,  1.08s/it]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- 24%|███████████▊                                     | 4399/18250 [2:01:34<5:01:45,  1.31s/it]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- 25%|████████████                                     | 4499/18250 [2:03:34<5:01:27,  1.32s/it]
- 25%|████████████                                     | 4500/18250 [2:03:36<4:57:07,  1.30s/it]The following columns in the evaluation set  don't have a corresponding argument in `Wav2Vec2ForCTC.forward` and have been ignored: input_length. If input_length are not expected by `Wav2Vec2ForCTC.forward`,  you can safely ignore this message.
-***** Running Evaluation *****
-  Num examples = 4843
-  Batch size = 8
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-Configuration saved in ./checkpoint-4500/config.json
-Model weights saved in ./checkpoint-4500/pytorch_model.bin
-Configuration saved in ./checkpoint-4500/preprocessor_config.json
-{'eval_loss': 0.35546526312828064, 'eval_wer': 0.3425205790519444, 'eval_runtime': 191.7841, 'eval_samples_per_second': 25.252, 'eval_steps_per_second': 3.16, 'epoch': 12.33}
-Configuration saved in ./preprocessor_config.json
-02/03/2022 00:38:21 - WARNING - huggingface_hub.repository - Several commits (4) will be pushed upstream.
-Several commits (4) will be pushed upstream.
-Deleting older checkpoint [checkpoint-3000] due to args.save_total_limit
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- 25%|████████████▎                                    | 4600/18250 [2:09:53<4:54:25,  1.29s/it]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- 26%|████████████▌                                    | 4699/18250 [2:11:52<4:54:26,  1.30s/it]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- 26%|████████████▉                                    | 4799/18250 [2:13:50<5:59:36,  1.60s/it]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- 27%|█████████████▏                                   | 4899/18250 [2:15:51<5:57:10,  1.61s/it]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- 27%|█████████████▍                                   | 5000/18250 [2:17:53<5:50:24,  1.59s/it]The following columns in the evaluation set  don't have a corresponding argument in `Wav2Vec2ForCTC.forward` and have been ignored: input_length. If input_length are not expected by `Wav2Vec2ForCTC.forward`,  you can safely ignore this message.
-***** Running Evaluation *****
-  Num examples = 4843
-  Batch size = 8
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-Configuration saved in ./checkpoint-5000/config.json
-{'eval_loss': 0.33404484391212463, 'eval_wer': 0.3227646891853534, 'eval_runtime': 191.5334, 'eval_samples_per_second': 25.285, 'eval_steps_per_second': 3.164, 'epoch': 13.7}
-Model weights saved in ./checkpoint-5000/pytorch_model.bin
-Configuration saved in ./checkpoint-5000/preprocessor_config.json
-Configuration saved in ./preprocessor_config.json
-Several commits (5) will be pushed upstream.
-02/03/2022 00:52:42 - WARNING - huggingface_hub.repository - Several commits (5) will be pushed upstream.
-Deleting older checkpoint [checkpoint-3500] due to args.save_total_limit
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- 28%|█████████████▋                                   | 5099/18250 [2:24:09<4:57:21,  1.36s/it]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- 28%|█████████████▉                                   | 5200/18250 [2:26:14<3:32:29,  1.02it/s]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- 29%|██████████████▏                                  | 5300/18250 [2:28:15<3:27:35,  1.04it/s]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- 30%|██████████████▌                                  | 5401/18250 [2:30:16<3:27:12,  1.03it/s]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
- 30%|██████████████▊                                  | 5499/18250 [2:32:17<4:24:05,  1.24s/it]
- 30%|██████████████▊                                  | 5500/18250 [2:32:18<4:19:58,  1.22s/it]The following columns in the evaluation set  don't have a corresponding argument in `Wav2Vec2ForCTC.forward` and have been ignored: input_length. If input_length are not expected by `Wav2Vec2ForCTC.forward`,  you can safely ignore this message.
-***** Running Evaluation *****
-  Num examples = 4843
-  Batch size = 8
-  2%|█▎                                                       | 14/606 [00:03<02:41,  3.67it/s]
-  4%|██                                                       | 22/606 [00:05<02:39,  3.66it/s]
-  5%|██▊                                                      | 30/606 [00:07<02:38,  3.62it/s]
-  6%|███▍                                                     | 37/606 [00:09<02:28,  3.84it/s]
-  7%|████▏                                                    | 44/606 [00:11<02:46,  3.37it/s]
-  9%|████▉                                                    | 52/606 [00:13<02:34,  3.60it/s]
- 10%|█████▋                                                   | 61/606 [00:15<02:06,  4.32it/s]
- 11%|██████▍                                                  | 68/606 [00:17<02:14,  4.01it/s]
- 12%|███████                                                  | 75/606 [00:19<02:43,  3.24it/s]
- 14%|███████▋                                                 | 82/606 [00:21<02:34,  3.38it/s]
- 15%|████████▎                                                | 88/606 [00:23<02:23,  3.60it/s]
- 16%|█████████                                                | 97/606 [00:25<02:17,  3.71it/s]
- 17%|█████████▌                                              | 104/606 [00:27<02:01,  4.12it/s]
- 19%|██████████▍                                             | 113/606 [00:29<01:56,  4.22it/s]
- 20%|███████████▎                                            | 122/606 [00:31<02:17,  3.53it/s]
- 21%|███████████▋                                            | 127/606 [00:33<02:14,  3.56it/s]
- 22%|████████████▍                                           | 135/606 [00:35<02:04,  3.79it/s]
- 24%|█████████████▎                                          | 144/606 [00:37<01:37,  4.74it/s]
- 25%|██████████████▏                                         | 153/606 [00:39<01:28,  5.13it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'eval_loss': 0.3372970223426819, 'eval_wer': 0.32949191030371844, 'eval_runtime': 191.1166, 'eval_samples_per_second': 25.341, 'eval_steps_per_second': 3.171, 'epoch': 15.07}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-02/03/2022 01:07:08 - WARNING - huggingface_hub.repository - Several commits (6) will be pushed upstream.
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.3158, 'learning_rate': 5.840307692307692e-05, 'epoch': 15.34}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.3406, 'learning_rate': 5.794153846153845e-05, 'epoch': 15.62}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.3312, 'learning_rate': 5.747999999999999e-05, 'epoch': 15.89}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.3263, 'learning_rate': 5.701846153846153e-05, 'epoch': 16.16}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.3055, 'learning_rate': 5.655692307692308e-05, 'epoch': 16.44}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|█████████���█████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'eval_loss': 0.3432324528694153, 'eval_wer': 0.3210332103321033, 'eval_runtime': 190.6681, 'eval_samples_per_second': 25.4, 'eval_steps_per_second': 3.178, 'epoch': 16.44}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.3106, 'learning_rate': 5.609538461538461e-05, 'epoch': 16.71}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████��                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.299, 'learning_rate': 5.563384615384615e-05, 'epoch': 16.98}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.2942, 'learning_rate': 5.517230769230769e-05, 'epoch': 17.26}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|████████████��██▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.2824, 'learning_rate': 5.471076923076922e-05, 'epoch': 17.53}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.3048, 'learning_rate': 5.4249230769230766e-05, 'epoch': 17.81}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'eval_loss': 0.3281940519809723, 'eval_wer': 0.3117797331819472, 'eval_runtime': 190.4965, 'eval_samples_per_second': 25.423, 'eval_steps_per_second': 3.181, 'epoch': 17.81}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|█████████��█████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.2909, 'learning_rate': 5.3787692307692306e-05, 'epoch': 18.08}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|██████████████���▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.2934, 'learning_rate': 5.332615384615384e-05, 'epoch': 18.36}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.2826, 'learning_rate': 5.286461538461538e-05, 'epoch': 18.63}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.2683, 'learning_rate': 5.240307692307692e-05, 'epoch': 18.9}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.2863, 'learning_rate': 5.194153846153846e-05, 'epoch': 19.18}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'eval_loss': 0.32255104184150696, 'eval_wer': 0.3018166335509509, 'eval_runtime': 191.9877, 'eval_samples_per_second': 25.226, 'eval_steps_per_second': 3.156, 'epoch': 19.18}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.2779, 'learning_rate': 5.1479999999999995e-05, 'epoch': 19.45}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.2452, 'learning_rate': 5.1018461538461534e-05, 'epoch': 19.73}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.2789, 'learning_rate': 5.055692307692307e-05, 'epoch': 20.0}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.2589, 'learning_rate': 5.009538461538461e-05, 'epoch': 20.27}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.2389, 'learning_rate': 4.963846153846154e-05, 'epoch': 20.55}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|█████████���█████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'eval_loss': 0.3049624264240265, 'eval_wer': 0.2985807550383196, 'eval_runtime': 191.4537, 'eval_samples_per_second': 25.296, 'eval_steps_per_second': 3.165, 'epoch': 20.55}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.2441, 'learning_rate': 4.917692307692307e-05, 'epoch': 20.82}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.2391, 'learning_rate': 4.8719999999999994e-05, 'epoch': 21.1}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|██████��████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.2227, 'learning_rate': 4.825846153846154e-05, 'epoch': 21.37}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████��███▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.2377, 'learning_rate': 4.779692307692307e-05, 'epoch': 21.64}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███���███████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.2361, 'learning_rate': 4.734e-05, 'epoch': 21.92}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|████████████���██▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'eval_loss': 0.30481210350990295, 'eval_wer': 0.29804144195288107, 'eval_runtime': 190.3668, 'eval_samples_per_second': 25.44, 'eval_steps_per_second': 3.183, 'epoch': 21.92}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|████████████��██▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.2605, 'learning_rate': 4.687846153846153e-05, 'epoch': 22.19}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|████���██████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.2185, 'learning_rate': 4.6416923076923076e-05, 'epoch': 22.47}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.2321, 'learning_rate': 4.5955384615384615e-05, 'epoch': 22.74}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|���██████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.2674, 'learning_rate': 4.549384615384615e-05, 'epoch': 23.01}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|█████���█████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.2263, 'learning_rate': 4.503230769230769e-05, 'epoch': 23.29}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'eval_loss': 0.3011174499988556, 'eval_wer': 0.2977008231620778, 'eval_runtime': 190.5, 'eval_samples_per_second': 25.423, 'eval_steps_per_second': 3.181, 'epoch': 23.29}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.2177, 'learning_rate': 4.4570769230769226e-05, 'epoch': 23.56}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.2375, 'learning_rate': 4.410923076923077e-05, 'epoch': 23.83}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.2119, 'learning_rate': 4.3647692307692304e-05, 'epoch': 24.11}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.221, 'learning_rate': 4.3186153846153844e-05, 'epoch': 24.38}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.2225, 'learning_rate': 4.2724615384615376e-05, 'epoch': 24.66}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|████████████���██▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'eval_loss': 0.3017021119594574, 'eval_wer': 0.2958841896111269, 'eval_runtime': 191.4744, 'eval_samples_per_second': 25.293, 'eval_steps_per_second': 3.165, 'epoch': 24.66}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.2195, 'learning_rate': 4.2263076923076915e-05, 'epoch': 24.93}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.2156, 'learning_rate': 4.180153846153846e-05, 'epoch': 25.21}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|█████████���█████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.1914, 'learning_rate': 4.134e-05, 'epoch': 25.48}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.1992, 'learning_rate': 4.087846153846153e-05, 'epoch': 25.75}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.2044, 'learning_rate': 4.041692307692307e-05, 'epoch': 26.03}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|█��█████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'eval_loss': 0.29767629504203796, 'eval_wer': 0.2782287822878229, 'eval_runtime': 190.7997, 'eval_samples_per_second': 25.383, 'eval_steps_per_second': 3.176, 'epoch': 26.03}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████���                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.1957, 'learning_rate': 3.995538461538461e-05, 'epoch': 26.3}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████���███████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.184, 'learning_rate': 3.949384615384615e-05, 'epoch': 26.57}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|████████████��██▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.1873, 'learning_rate': 3.903230769230769e-05, 'epoch': 26.85}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|████���██████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.1762, 'learning_rate': 3.857076923076923e-05, 'epoch': 27.12}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|█████████���█████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.2017, 'learning_rate': 3.810923076923076e-05, 'epoch': 27.4}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|█���█████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'eval_loss': 0.29657357931137085, 'eval_wer': 0.27814362759012207, 'eval_runtime': 191.6103, 'eval_samples_per_second': 25.275, 'eval_steps_per_second': 3.163, 'epoch': 27.4}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|█��█████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.1837, 'learning_rate': 3.76476923076923e-05, 'epoch': 27.67}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.1948, 'learning_rate': 3.718615384615384e-05, 'epoch': 27.94}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.1852, 'learning_rate': 3.6724615384615386e-05, 'epoch': 28.22}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.1681, 'learning_rate': 3.626307692307692e-05, 'epoch': 28.49}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.1912, 'learning_rate': 3.580153846153846e-05, 'epoch': 28.77}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'eval_loss': 0.2999463379383087, 'eval_wer': 0.2785977859778598, 'eval_runtime': 192.5233, 'eval_samples_per_second': 25.155, 'eval_steps_per_second': 3.148, 'epoch': 28.77}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.1658, 'learning_rate': 3.534e-05, 'epoch': 29.04}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.1691, 'learning_rate': 3.4878461538461536e-05, 'epoch': 29.31}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.1531, 'learning_rate': 3.4416923076923076e-05, 'epoch': 29.59}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.1569, 'learning_rate': 3.3955384615384615e-05, 'epoch': 29.86}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|█████████████���█▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.1658, 'learning_rate': 3.349384615384615e-05, 'epoch': 30.14}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'eval_loss': 0.29912158846855164, 'eval_wer': 0.2757309111552654, 'eval_runtime': 192.0401, 'eval_samples_per_second': 25.219, 'eval_steps_per_second': 3.156, 'epoch': 30.14}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|█████���█████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.1565, 'learning_rate': 3.303230769230769e-05, 'epoch': 30.41}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|██████████���████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.1556, 'learning_rate': 3.2570769230769226e-05, 'epoch': 30.68}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|█��█████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.1538, 'learning_rate': 3.2109230769230765e-05, 'epoch': 30.96}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|██████���████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.1531, 'learning_rate': 3.1647692307692304e-05, 'epoch': 31.23}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.148, 'learning_rate': 3.118615384615384e-05, 'epoch': 31.51}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'eval_loss': 0.2915021479129791, 'eval_wer': 0.268379222253761, 'eval_runtime': 191.8689, 'eval_samples_per_second': 25.241, 'eval_steps_per_second': 3.158, 'epoch': 31.51}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.1414, 'learning_rate': 3.072461538461538e-05, 'epoch': 31.78}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.1364, 'learning_rate': 3.026307692307692e-05, 'epoch': 32.05}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|��██████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.1531, 'learning_rate': 2.980153846153846e-05, 'epoch': 32.33}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.1331, 'learning_rate': 2.9339999999999997e-05, 'epoch': 32.6}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.1423, 'learning_rate': 2.8878461538461536e-05, 'epoch': 32.88}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'eval_loss': 0.29132989048957825, 'eval_wer': 0.2642634118648879, 'eval_runtime': 191.3127, 'eval_samples_per_second': 25.315, 'eval_steps_per_second': 3.168, 'epoch': 32.88}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.1527, 'learning_rate': 2.8416923076923075e-05, 'epoch': 33.15}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.1155, 'learning_rate': 2.795538461538461e-05, 'epoch': 33.42}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.1082, 'learning_rate': 2.7498461538461536e-05, 'epoch': 33.7}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.1172, 'learning_rate': 2.703692307692307e-05, 'epoch': 33.97}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████��███▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.123, 'learning_rate': 2.6575384615384614e-05, 'epoch': 34.25}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███���███████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'eval_loss': 0.2776812016963959, 'eval_wer': 0.26304286119784276, 'eval_runtime': 191.5218, 'eval_samples_per_second': 25.287, 'eval_steps_per_second': 3.164, 'epoch': 34.25}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███��███████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.1189, 'learning_rate': 2.611384615384615e-05, 'epoch': 34.52}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|████████��██████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.1115, 'learning_rate': 2.5652307692307693e-05, 'epoch': 34.79}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|█████████████���█▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.1182, 'learning_rate': 2.519076923076923e-05, 'epoch': 35.07}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.1172, 'learning_rate': 2.4729230769230764e-05, 'epoch': 35.34}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|█████████��█████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.1297, 'learning_rate': 2.4267692307692307e-05, 'epoch': 35.62}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'eval_loss': 0.28734585642814636, 'eval_wer': 0.26460403065569116, 'eval_runtime': 191.3152, 'eval_samples_per_second': 25.314, 'eval_steps_per_second': 3.168, 'epoch': 35.62}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.1112, 'learning_rate': 2.3806153846153843e-05, 'epoch': 35.89}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.1157, 'learning_rate': 2.3349230769230767e-05, 'epoch': 36.16}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.0938, 'learning_rate': 2.2887692307692303e-05, 'epoch': 36.44}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.1007, 'learning_rate': 2.2426153846153846e-05, 'epoch': 36.71}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.0987, 'learning_rate': 2.1964615384615382e-05, 'epoch': 36.98}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'eval_loss': 0.28285086154937744, 'eval_wer': 0.2619358501277321, 'eval_runtime': 190.4602, 'eval_samples_per_second': 25.428, 'eval_steps_per_second': 3.182, 'epoch': 36.98}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|██████████████���▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.0967, 'learning_rate': 2.1503076923076924e-05, 'epoch': 37.26}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.0922, 'learning_rate': 2.104153846153846e-05, 'epoch': 37.53}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|██████████��████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.0885, 'learning_rate': 2.0579999999999996e-05, 'epoch': 37.81}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████���                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.0979, 'learning_rate': 2.011846153846154e-05, 'epoch': 38.08}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.0873, 'learning_rate': 1.9656923076923075e-05, 'epoch': 38.36}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████��███▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'eval_loss': 0.2864430546760559, 'eval_wer': 0.26080045415838776, 'eval_runtime': 191.1374, 'eval_samples_per_second': 25.338, 'eval_steps_per_second': 3.17, 'epoch': 38.36}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|████████████���██▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.0868, 'learning_rate': 1.9195384615384614e-05, 'epoch': 38.63}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███��███████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.0831, 'learning_rate': 1.8733846153846153e-05, 'epoch': 38.9}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|████████��██████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.0907, 'learning_rate': 1.8272307692307692e-05, 'epoch': 39.18}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.093, 'learning_rate': 1.781076923076923e-05, 'epoch': 39.45}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.0848, 'learning_rate': 1.7349230769230767e-05, 'epoch': 39.73}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'eval_loss': 0.2827305793762207, 'eval_wer': 0.2576781152426909, 'eval_runtime': 193.1681, 'eval_samples_per_second': 25.071, 'eval_steps_per_second': 3.137, 'epoch': 39.73}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.0913, 'learning_rate': 1.6887692307692307e-05, 'epoch': 40.0}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.0764, 'learning_rate': 1.6426153846153846e-05, 'epoch': 40.27}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.0668, 'learning_rate': 1.5964615384615385e-05, 'epoch': 40.55}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|██████████████��▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.0845, 'learning_rate': 1.550307692307692e-05, 'epoch': 40.82}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.0628, 'learning_rate': 1.5041538461538462e-05, 'epoch': 41.1}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████���███▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'eval_loss': 0.28957271575927734, 'eval_wer': 0.258103888731195, 'eval_runtime': 192.7832, 'eval_samples_per_second': 25.121, 'eval_steps_per_second': 3.143, 'epoch': 41.1}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|████████████���██▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.0746, 'learning_rate': 1.4579999999999998e-05, 'epoch': 41.37}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███��███████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.0889, 'learning_rate': 1.4118461538461537e-05, 'epoch': 41.64}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|████████���██████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.0669, 'learning_rate': 1.3656923076923076e-05, 'epoch': 41.92}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.069, 'learning_rate': 1.3195384615384615e-05, 'epoch': 42.19}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|████��██████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.0815, 'learning_rate': 1.2733846153846153e-05, 'epoch': 42.47}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|█████████���█████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'eval_loss': 0.28144747018814087, 'eval_wer': 0.25608856088560883, 'eval_runtime': 192.799, 'eval_samples_per_second': 25.119, 'eval_steps_per_second': 3.143, 'epoch': 42.47}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.0468, 'learning_rate': 1.227230769230769e-05, 'epoch': 42.74}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|█���█████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.073, 'learning_rate': 1.1815384615384615e-05, 'epoch': 43.01}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.05, 'learning_rate': 1.1353846153846154e-05, 'epoch': 43.29}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.0562, 'learning_rate': 1.089230769230769e-05, 'epoch': 43.56}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.0587, 'learning_rate': 1.043076923076923e-05, 'epoch': 43.83}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████���███▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'eval_loss': 0.27383875846862793, 'eval_wer': 0.25421515753619073, 'eval_runtime': 191.944, 'eval_samples_per_second': 25.231, 'eval_steps_per_second': 3.157, 'epoch': 43.83}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.0479, 'learning_rate': 9.969230769230769e-06, 'epoch': 44.11}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.0464, 'learning_rate': 9.507692307692308e-06, 'epoch': 44.38}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.0515, 'learning_rate': 9.046153846153845e-06, 'epoch': 44.66}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|█████████████���█▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.0456, 'learning_rate': 8.584615384615385e-06, 'epoch': 44.93}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.0709, 'learning_rate': 8.123076923076922e-06, 'epoch': 45.21}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'eval_loss': 0.27852663397789, 'eval_wer': 0.25784842463809254, 'eval_runtime': 191.9755, 'eval_samples_per_second': 25.227, 'eval_steps_per_second': 3.157, 'epoch': 45.21}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.0616, 'learning_rate': 7.661538461538461e-06, 'epoch': 45.48}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|██��████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.0707, 'learning_rate': 7.2e-06, 'epoch': 45.75}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|████████████���██▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.037, 'learning_rate': 6.738461538461537e-06, 'epoch': 46.03}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|████���██████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.0475, 'learning_rate': 6.2769230769230764e-06, 'epoch': 46.3}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|█████████���█████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.0512, 'learning_rate': 5.815384615384615e-06, 'epoch': 46.57}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|██████████████���▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'eval_loss': 0.27928975224494934, 'eval_wer': 0.2539029236446211, 'eval_runtime': 191.7381, 'eval_samples_per_second': 25.258, 'eval_steps_per_second': 3.161, 'epoch': 46.57}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|█���█████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.0454, 'learning_rate': 5.353846153846153e-06, 'epoch': 46.85}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|██████���████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.0505, 'learning_rate': 4.892307692307692e-06, 'epoch': 47.12}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.0351, 'learning_rate': 4.43076923076923e-06, 'epoch': 47.4}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███��███████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.0591, 'learning_rate': 3.969230769230768e-06, 'epoch': 47.67}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.0396, 'learning_rate': 3.5076923076923076e-06, 'epoch': 47.94}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'eval_loss': 0.27876144647598267, 'eval_wer': 0.25251206358217426, 'eval_runtime': 191.8086, 'eval_samples_per_second': 25.249, 'eval_steps_per_second': 3.159, 'epoch': 47.94}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.0482, 'learning_rate': 3.046153846153846e-06, 'epoch': 48.22}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.0433, 'learning_rate': 2.5846153846153844e-06, 'epoch': 48.49}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.036, 'learning_rate': 2.1230769230769228e-06, 'epoch': 48.77}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.0739, 'learning_rate': 1.6615384615384613e-06, 'epoch': 49.04}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.0481, 'learning_rate': 1.2046153846153845e-06, 'epoch': 49.31}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|█████████████��█▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|���██████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'eval_loss': 0.27769750356674194, 'eval_wer': 0.2534487652568833, 'eval_runtime': 191.0585, 'eval_samples_per_second': 25.348, 'eval_steps_per_second': 3.172, 'epoch': 49.31}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.0195, 'learning_rate': 7.43076923076923e-07, 'epoch': 49.59}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'loss': 1.0473, 'learning_rate': 2.8153846153846154e-07, 'epoch': 49.86}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|██████████��████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-{'train_runtime': 31196.8756, 'train_samples_per_second': 18.736, 'train_steps_per_second': 0.585, 'train_loss': 1.449225697974636, 'epoch': 50.0}
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████���███████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
- 27%|███████████████▏                                        | 164/606 [00:41<01:41,  4.34it/s]
-Upload file runs/Feb02_22-30-13_brutasse/events.out.tfevents.1643841025.brutasse.21084.0: 100%|
-Upload file runs/Feb02_22-30-13_brutasse/events.out.tfevents.1643841025.brutasse.21084.0: 100%|
-02/03/2022 07:15:24 - WARNING - huggingface_hub.repository - To https://huggingface.co/patrickvonplaten/xls-r-300m-sv-cv8
-   076115a..b03bb0b  main -> main
-Upload file runs/Feb02_22-30-13_brutasse/events.out.tfevents.1643841025.brutasse.21084.0: 100%|
-Upload file runs/Feb02_22-30-13_brutasse/events.out.tfevents.1643841025.brutasse.21084.0: 100%|
-Upload file runs/Feb02_22-30-13_brutasse/events.out.tfevents.1643841025.brutasse.21084.0: 100%|
-Upload file runs/Feb02_22-30-13_brutasse/events.out.tfevents.1643841025.brutasse.21084.0: 100%|
-Upload file runs/Feb02_22-30-13_brutasse/events.out.tfevents.1643841025.brutasse.21084.0: 100%|
-Upload file runs/Feb02_22-30-13_brutasse/events.out.tfevents.1643841025.brutasse.21084.0: 100%|
-Upload file runs/Feb02_22-30-13_brutasse/events.out.tfevents.1643841025.brutasse.21084.0: 100%|
-Upload file runs/Feb02_22-30-13_brutasse/events.out.tfevents.1643841025.brutasse.21084.0: 100%|
-Upload file runs/Feb02_22-30-13_brutasse/events.out.tfevents.1643841025.brutasse.21084.0: 100%|
-Upload file runs/Feb02_22-30-13_brutasse/events.out.tfevents.1643841025.brutasse.21084.0: 100%|
-Upload file runs/Feb02_22-30-13_brutasse/events.out.tfevents.1643841025.brutasse.21084.0: 100%|
-02/03/2022 07:15:49 - WARNING - huggingface_hub.repository - To https://huggingface.co/patrickvonplaten/xls-r-300m-sv-cv8
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-***** train metrics *****
-  epoch                    =       50.0
-  train_loss               =     1.4492
-  train_runtime            = 8:39:56.87
-  train_samples            =      11690
-  train_samples_per_second =     18.736
-  train_steps_per_second   =      0.585
-02/03/2022 07:15:52 - INFO - __main__ - *** Evaluate ***
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-***** eval metrics *****
-  epoch                   =       50.0
-  eval_loss               =     0.2779
-  eval_runtime            = 0:03:10.08
-  eval_samples            =       4843
-  eval_samples_per_second =     25.478
-  eval_steps_per_second   =      3.188
-  eval_wer                =     0.2525
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-  0%|                                                                  | 0/606 [00:00<?, ?it/s]
-Upload file runs/Feb02_22-30-13_brutasse/events.out.tfevents.1643872742.brutasse.21084.2: 100%|
-Upload file runs/Feb02_22-30-13_brutasse/events.out.tfevents.1643872742.brutasse.21084.2: 100%|
-02/03/2022 07:19:47 - WARNING - huggingface_hub.repository - To https://huggingface.co/patrickvonplaten/xls-r-300m-sv-cv8