Training in progress, epoch 55, checkpoint
Browse files- checkpoint-28215/config.json +82 -0
- checkpoint-28215/model.safetensors +3 -0
- checkpoint-28215/optimizer.pt +3 -0
- checkpoint-28215/preprocessor_config.json +11 -0
- checkpoint-28215/rng_state.pth +3 -0
- checkpoint-28215/scheduler.pt +3 -0
- checkpoint-28215/trainer_state.json +977 -0
- checkpoint-28215/training_args.bin +3 -0
checkpoint-28215/config.json
ADDED
|
@@ -0,0 +1,82 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "facebook/w2v-bert-2.0",
|
| 3 |
+
"activation_dropout": 0.0,
|
| 4 |
+
"adapter_act": "relu",
|
| 5 |
+
"adapter_kernel_size": 3,
|
| 6 |
+
"adapter_stride": 2,
|
| 7 |
+
"add_adapter": true,
|
| 8 |
+
"apply_spec_augment": false,
|
| 9 |
+
"architectures": [
|
| 10 |
+
"Wav2Vec2BertForCTC"
|
| 11 |
+
],
|
| 12 |
+
"attention_dropout": 0.0,
|
| 13 |
+
"bos_token_id": 1,
|
| 14 |
+
"classifier_proj_size": 768,
|
| 15 |
+
"codevector_dim": 768,
|
| 16 |
+
"conformer_conv_dropout": 0.1,
|
| 17 |
+
"contrastive_logits_temperature": 0.1,
|
| 18 |
+
"conv_depthwise_kernel_size": 31,
|
| 19 |
+
"ctc_loss_reduction": "mean",
|
| 20 |
+
"ctc_zero_infinity": true,
|
| 21 |
+
"diversity_loss_weight": 0.1,
|
| 22 |
+
"eos_token_id": 2,
|
| 23 |
+
"feat_proj_dropout": 0.0,
|
| 24 |
+
"feat_quantizer_dropout": 0.0,
|
| 25 |
+
"feature_projection_input_dim": 160,
|
| 26 |
+
"final_dropout": 0.1,
|
| 27 |
+
"hidden_act": "swish",
|
| 28 |
+
"hidden_dropout": 0.0,
|
| 29 |
+
"hidden_size": 1024,
|
| 30 |
+
"initializer_range": 0.02,
|
| 31 |
+
"intermediate_size": 4096,
|
| 32 |
+
"layer_norm_eps": 1e-05,
|
| 33 |
+
"layerdrop": 0.0,
|
| 34 |
+
"left_max_position_embeddings": 64,
|
| 35 |
+
"mask_feature_length": 10,
|
| 36 |
+
"mask_feature_min_masks": 0,
|
| 37 |
+
"mask_feature_prob": 0.0,
|
| 38 |
+
"mask_time_length": 10,
|
| 39 |
+
"mask_time_min_masks": 2,
|
| 40 |
+
"mask_time_prob": 0.0,
|
| 41 |
+
"max_source_positions": 5000,
|
| 42 |
+
"model_type": "wav2vec2-bert",
|
| 43 |
+
"num_adapter_layers": 1,
|
| 44 |
+
"num_attention_heads": 16,
|
| 45 |
+
"num_codevector_groups": 2,
|
| 46 |
+
"num_codevectors_per_group": 320,
|
| 47 |
+
"num_hidden_layers": 24,
|
| 48 |
+
"num_negatives": 100,
|
| 49 |
+
"output_hidden_size": 1024,
|
| 50 |
+
"pad_token_id": 45,
|
| 51 |
+
"position_embeddings_type": "relative_key",
|
| 52 |
+
"proj_codevector_dim": 768,
|
| 53 |
+
"right_max_position_embeddings": 8,
|
| 54 |
+
"rotary_embedding_base": 10000,
|
| 55 |
+
"tdnn_dilation": [
|
| 56 |
+
1,
|
| 57 |
+
2,
|
| 58 |
+
3,
|
| 59 |
+
1,
|
| 60 |
+
1
|
| 61 |
+
],
|
| 62 |
+
"tdnn_dim": [
|
| 63 |
+
512,
|
| 64 |
+
512,
|
| 65 |
+
512,
|
| 66 |
+
512,
|
| 67 |
+
1500
|
| 68 |
+
],
|
| 69 |
+
"tdnn_kernel": [
|
| 70 |
+
5,
|
| 71 |
+
3,
|
| 72 |
+
3,
|
| 73 |
+
1,
|
| 74 |
+
1
|
| 75 |
+
],
|
| 76 |
+
"torch_dtype": "float32",
|
| 77 |
+
"transformers_version": "4.46.0",
|
| 78 |
+
"use_intermediate_ffn_before_adapter": false,
|
| 79 |
+
"use_weighted_layer_sum": false,
|
| 80 |
+
"vocab_size": 48,
|
| 81 |
+
"xvector_output_dim": 512
|
| 82 |
+
}
|
checkpoint-28215/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b250c8f7c7f3dd1d7138c728b64a32a474aec18d499e997ebbe96f9018ce7239
|
| 3 |
+
size 2423011360
|
checkpoint-28215/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1c80f29d358417f63da9c31e12136927c6b6cadf98375d2dfc2674c1012c9c73
|
| 3 |
+
size 4846483562
|
checkpoint-28215/preprocessor_config.json
ADDED
|
@@ -0,0 +1,11 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"feature_extractor_type": "SeamlessM4TFeatureExtractor",
|
| 3 |
+
"feature_size": 80,
|
| 4 |
+
"num_mel_bins": 80,
|
| 5 |
+
"padding_side": "right",
|
| 6 |
+
"padding_value": 1,
|
| 7 |
+
"processor_class": "Wav2Vec2BertProcessor",
|
| 8 |
+
"return_attention_mask": true,
|
| 9 |
+
"sampling_rate": 16000,
|
| 10 |
+
"stride": 2
|
| 11 |
+
}
|
checkpoint-28215/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:80009179e148820d7fb270929a4aecdb0326bfbaba7fe12b1ee1778f3e309a00
|
| 3 |
+
size 14244
|
checkpoint-28215/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d27c732d1436d798a0f3ea75dba2fe8baf5de7dfd47462bac4b4a00c5df097da
|
| 3 |
+
size 1064
|
checkpoint-28215/trainer_state.json
ADDED
|
@@ -0,0 +1,977 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.3606263447286636,
|
| 3 |
+
"best_model_checkpoint": "w2v-bert-2.0-CV_Fleurs-lg-10hrs-v4/checkpoint-25137",
|
| 4 |
+
"epoch": 55.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 28215,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 1.0,
|
| 13 |
+
"grad_norm": 30.792112350463867,
|
| 14 |
+
"learning_rate": 9.998031496062993e-05,
|
| 15 |
+
"loss": 1.4106,
|
| 16 |
+
"step": 513
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 1.0,
|
| 20 |
+
"eval_cer": 0.1264185884550246,
|
| 21 |
+
"eval_loss": 0.5293365716934204,
|
| 22 |
+
"eval_runtime": 31.0205,
|
| 23 |
+
"eval_samples_per_second": 26.628,
|
| 24 |
+
"eval_steps_per_second": 6.673,
|
| 25 |
+
"eval_wer": 0.5665790102797036,
|
| 26 |
+
"step": 513
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 2.0,
|
| 30 |
+
"grad_norm": 10.16055679321289,
|
| 31 |
+
"learning_rate": 9.897047244094488e-05,
|
| 32 |
+
"loss": 0.4775,
|
| 33 |
+
"step": 1026
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 2.0,
|
| 37 |
+
"eval_cer": 0.11182513176846737,
|
| 38 |
+
"eval_loss": 0.4882170259952545,
|
| 39 |
+
"eval_runtime": 30.4398,
|
| 40 |
+
"eval_samples_per_second": 27.135,
|
| 41 |
+
"eval_steps_per_second": 6.8,
|
| 42 |
+
"eval_wer": 0.5251016017212526,
|
| 43 |
+
"step": 1026
|
| 44 |
+
},
|
| 45 |
+
{
|
| 46 |
+
"epoch": 3.0,
|
| 47 |
+
"grad_norm": 24.60751724243164,
|
| 48 |
+
"learning_rate": 9.796062992125985e-05,
|
| 49 |
+
"loss": 0.3696,
|
| 50 |
+
"step": 1539
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 3.0,
|
| 54 |
+
"eval_cer": 0.09981164627486981,
|
| 55 |
+
"eval_loss": 0.397329181432724,
|
| 56 |
+
"eval_runtime": 30.5019,
|
| 57 |
+
"eval_samples_per_second": 27.08,
|
| 58 |
+
"eval_steps_per_second": 6.786,
|
| 59 |
+
"eval_wer": 0.4564905570164953,
|
| 60 |
+
"step": 1539
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"epoch": 4.0,
|
| 64 |
+
"grad_norm": 8.973501205444336,
|
| 65 |
+
"learning_rate": 9.695078740157481e-05,
|
| 66 |
+
"loss": 0.3007,
|
| 67 |
+
"step": 2052
|
| 68 |
+
},
|
| 69 |
+
{
|
| 70 |
+
"epoch": 4.0,
|
| 71 |
+
"eval_cer": 0.09322717991737761,
|
| 72 |
+
"eval_loss": 0.3895888030529022,
|
| 73 |
+
"eval_runtime": 29.8209,
|
| 74 |
+
"eval_samples_per_second": 27.699,
|
| 75 |
+
"eval_steps_per_second": 6.941,
|
| 76 |
+
"eval_wer": 0.4311498924217069,
|
| 77 |
+
"step": 2052
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 5.0,
|
| 81 |
+
"grad_norm": 3.461097240447998,
|
| 82 |
+
"learning_rate": 9.594094488188978e-05,
|
| 83 |
+
"loss": 0.2564,
|
| 84 |
+
"step": 2565
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 5.0,
|
| 88 |
+
"eval_cer": 0.09659855331676664,
|
| 89 |
+
"eval_loss": 0.3743041455745697,
|
| 90 |
+
"eval_runtime": 29.9644,
|
| 91 |
+
"eval_samples_per_second": 27.566,
|
| 92 |
+
"eval_steps_per_second": 6.908,
|
| 93 |
+
"eval_wer": 0.4372459956968683,
|
| 94 |
+
"step": 2565
|
| 95 |
+
},
|
| 96 |
+
{
|
| 97 |
+
"epoch": 6.0,
|
| 98 |
+
"grad_norm": 9.60593032836914,
|
| 99 |
+
"learning_rate": 9.493110236220472e-05,
|
| 100 |
+
"loss": 0.2273,
|
| 101 |
+
"step": 3078
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 6.0,
|
| 105 |
+
"eval_cer": 0.0879880973108153,
|
| 106 |
+
"eval_loss": 0.3549309968948364,
|
| 107 |
+
"eval_runtime": 30.2119,
|
| 108 |
+
"eval_samples_per_second": 27.34,
|
| 109 |
+
"eval_steps_per_second": 6.852,
|
| 110 |
+
"eval_wer": 0.41836002868754485,
|
| 111 |
+
"step": 3078
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 7.0,
|
| 115 |
+
"grad_norm": 2.7023348808288574,
|
| 116 |
+
"learning_rate": 9.392125984251969e-05,
|
| 117 |
+
"loss": 0.203,
|
| 118 |
+
"step": 3591
|
| 119 |
+
},
|
| 120 |
+
{
|
| 121 |
+
"epoch": 7.0,
|
| 122 |
+
"eval_cer": 0.08468003608794061,
|
| 123 |
+
"eval_loss": 0.3835783302783966,
|
| 124 |
+
"eval_runtime": 29.6315,
|
| 125 |
+
"eval_samples_per_second": 27.876,
|
| 126 |
+
"eval_steps_per_second": 6.986,
|
| 127 |
+
"eval_wer": 0.4000717188620607,
|
| 128 |
+
"step": 3591
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 8.0,
|
| 132 |
+
"grad_norm": 2.196669340133667,
|
| 133 |
+
"learning_rate": 9.291141732283465e-05,
|
| 134 |
+
"loss": 0.1775,
|
| 135 |
+
"step": 4104
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 8.0,
|
| 139 |
+
"eval_cer": 0.08550309438262714,
|
| 140 |
+
"eval_loss": 0.3693934977054596,
|
| 141 |
+
"eval_runtime": 30.2095,
|
| 142 |
+
"eval_samples_per_second": 27.342,
|
| 143 |
+
"eval_steps_per_second": 6.852,
|
| 144 |
+
"eval_wer": 0.4132201769065264,
|
| 145 |
+
"step": 4104
|
| 146 |
+
},
|
| 147 |
+
{
|
| 148 |
+
"epoch": 9.0,
|
| 149 |
+
"grad_norm": 3.045036792755127,
|
| 150 |
+
"learning_rate": 9.190354330708662e-05,
|
| 151 |
+
"loss": 0.159,
|
| 152 |
+
"step": 4617
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 9.0,
|
| 156 |
+
"eval_cer": 0.0889694360467877,
|
| 157 |
+
"eval_loss": 0.3487217128276825,
|
| 158 |
+
"eval_runtime": 29.9136,
|
| 159 |
+
"eval_samples_per_second": 27.613,
|
| 160 |
+
"eval_steps_per_second": 6.92,
|
| 161 |
+
"eval_wer": 0.4136983026535979,
|
| 162 |
+
"step": 4617
|
| 163 |
+
},
|
| 164 |
+
{
|
| 165 |
+
"epoch": 10.0,
|
| 166 |
+
"grad_norm": 3.3439998626708984,
|
| 167 |
+
"learning_rate": 9.089370078740158e-05,
|
| 168 |
+
"loss": 0.1444,
|
| 169 |
+
"step": 5130
|
| 170 |
+
},
|
| 171 |
+
{
|
| 172 |
+
"epoch": 10.0,
|
| 173 |
+
"eval_cer": 0.08915937257633075,
|
| 174 |
+
"eval_loss": 0.3705226480960846,
|
| 175 |
+
"eval_runtime": 30.2455,
|
| 176 |
+
"eval_samples_per_second": 27.31,
|
| 177 |
+
"eval_steps_per_second": 6.844,
|
| 178 |
+
"eval_wer": 0.42768348075543866,
|
| 179 |
+
"step": 5130
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 11.0,
|
| 183 |
+
"grad_norm": 3.339082956314087,
|
| 184 |
+
"learning_rate": 8.988385826771653e-05,
|
| 185 |
+
"loss": 0.1262,
|
| 186 |
+
"step": 5643
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 11.0,
|
| 190 |
+
"eval_cer": 0.08284398296902452,
|
| 191 |
+
"eval_loss": 0.3619905114173889,
|
| 192 |
+
"eval_runtime": 30.7416,
|
| 193 |
+
"eval_samples_per_second": 26.869,
|
| 194 |
+
"eval_steps_per_second": 6.734,
|
| 195 |
+
"eval_wer": 0.3981592158737748,
|
| 196 |
+
"step": 5643
|
| 197 |
+
},
|
| 198 |
+
{
|
| 199 |
+
"epoch": 12.0,
|
| 200 |
+
"grad_norm": 0.1660577654838562,
|
| 201 |
+
"learning_rate": 8.88740157480315e-05,
|
| 202 |
+
"loss": 0.1139,
|
| 203 |
+
"step": 6156
|
| 204 |
+
},
|
| 205 |
+
{
|
| 206 |
+
"epoch": 12.0,
|
| 207 |
+
"eval_cer": 0.08422102280821159,
|
| 208 |
+
"eval_loss": 0.3665634095668793,
|
| 209 |
+
"eval_runtime": 30.3996,
|
| 210 |
+
"eval_samples_per_second": 27.171,
|
| 211 |
+
"eval_steps_per_second": 6.809,
|
| 212 |
+
"eval_wer": 0.40545063351661487,
|
| 213 |
+
"step": 6156
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 13.0,
|
| 217 |
+
"grad_norm": 0.19767990708351135,
|
| 218 |
+
"learning_rate": 8.786417322834646e-05,
|
| 219 |
+
"loss": 0.1015,
|
| 220 |
+
"step": 6669
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"epoch": 13.0,
|
| 224 |
+
"eval_cer": 0.08355624495481094,
|
| 225 |
+
"eval_loss": 0.3656393885612488,
|
| 226 |
+
"eval_runtime": 30.0647,
|
| 227 |
+
"eval_samples_per_second": 27.474,
|
| 228 |
+
"eval_steps_per_second": 6.885,
|
| 229 |
+
"eval_wer": 0.40114750179297154,
|
| 230 |
+
"step": 6669
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 14.0,
|
| 234 |
+
"grad_norm": 2.6208112239837646,
|
| 235 |
+
"learning_rate": 8.685433070866143e-05,
|
| 236 |
+
"loss": 0.0914,
|
| 237 |
+
"step": 7182
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 14.0,
|
| 241 |
+
"eval_cer": 0.08339796451352506,
|
| 242 |
+
"eval_loss": 0.3572923541069031,
|
| 243 |
+
"eval_runtime": 30.3516,
|
| 244 |
+
"eval_samples_per_second": 27.214,
|
| 245 |
+
"eval_steps_per_second": 6.82,
|
| 246 |
+
"eval_wer": 0.39278030121922064,
|
| 247 |
+
"step": 7182
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 15.0,
|
| 251 |
+
"grad_norm": 3.129692316055298,
|
| 252 |
+
"learning_rate": 8.584448818897639e-05,
|
| 253 |
+
"loss": 0.0836,
|
| 254 |
+
"step": 7695
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 15.0,
|
| 258 |
+
"eval_cer": 0.07895028411339211,
|
| 259 |
+
"eval_loss": 0.379753053188324,
|
| 260 |
+
"eval_runtime": 30.1224,
|
| 261 |
+
"eval_samples_per_second": 27.421,
|
| 262 |
+
"eval_steps_per_second": 6.872,
|
| 263 |
+
"eval_wer": 0.3795123117379871,
|
| 264 |
+
"step": 7695
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 16.0,
|
| 268 |
+
"grad_norm": 4.029957294464111,
|
| 269 |
+
"learning_rate": 8.483464566929134e-05,
|
| 270 |
+
"loss": 0.0729,
|
| 271 |
+
"step": 8208
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 16.0,
|
| 275 |
+
"eval_cer": 0.08219503315975245,
|
| 276 |
+
"eval_loss": 0.4288513660430908,
|
| 277 |
+
"eval_runtime": 30.1728,
|
| 278 |
+
"eval_samples_per_second": 27.376,
|
| 279 |
+
"eval_steps_per_second": 6.86,
|
| 280 |
+
"eval_wer": 0.3895529524264882,
|
| 281 |
+
"step": 8208
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 17.0,
|
| 285 |
+
"grad_norm": 0.5638870596885681,
|
| 286 |
+
"learning_rate": 8.38248031496063e-05,
|
| 287 |
+
"loss": 0.0652,
|
| 288 |
+
"step": 8721
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 17.0,
|
| 292 |
+
"eval_cer": 0.087386631633929,
|
| 293 |
+
"eval_loss": 0.4778367280960083,
|
| 294 |
+
"eval_runtime": 30.6189,
|
| 295 |
+
"eval_samples_per_second": 26.977,
|
| 296 |
+
"eval_steps_per_second": 6.761,
|
| 297 |
+
"eval_wer": 0.4157303370786517,
|
| 298 |
+
"step": 8721
|
| 299 |
+
},
|
| 300 |
+
{
|
| 301 |
+
"epoch": 18.0,
|
| 302 |
+
"grad_norm": 0.010239621624350548,
|
| 303 |
+
"learning_rate": 8.281496062992127e-05,
|
| 304 |
+
"loss": 0.0599,
|
| 305 |
+
"step": 9234
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 18.0,
|
| 309 |
+
"eval_cer": 0.07918770477532092,
|
| 310 |
+
"eval_loss": 0.4570063352584839,
|
| 311 |
+
"eval_runtime": 30.2821,
|
| 312 |
+
"eval_samples_per_second": 27.277,
|
| 313 |
+
"eval_steps_per_second": 6.836,
|
| 314 |
+
"eval_wer": 0.37963184317475496,
|
| 315 |
+
"step": 9234
|
| 316 |
+
},
|
| 317 |
+
{
|
| 318 |
+
"epoch": 19.0,
|
| 319 |
+
"grad_norm": 3.2985517978668213,
|
| 320 |
+
"learning_rate": 8.180511811023623e-05,
|
| 321 |
+
"loss": 0.0504,
|
| 322 |
+
"step": 9747
|
| 323 |
+
},
|
| 324 |
+
{
|
| 325 |
+
"epoch": 19.0,
|
| 326 |
+
"eval_cer": 0.08564554677978442,
|
| 327 |
+
"eval_loss": 0.44442683458328247,
|
| 328 |
+
"eval_runtime": 30.5951,
|
| 329 |
+
"eval_samples_per_second": 26.998,
|
| 330 |
+
"eval_steps_per_second": 6.766,
|
| 331 |
+
"eval_wer": 0.4125029882859192,
|
| 332 |
+
"step": 9747
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 20.0,
|
| 336 |
+
"grad_norm": 0.014229847118258476,
|
| 337 |
+
"learning_rate": 8.079724409448819e-05,
|
| 338 |
+
"loss": 0.0498,
|
| 339 |
+
"step": 10260
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 20.0,
|
| 343 |
+
"eval_cer": 0.08422102280821159,
|
| 344 |
+
"eval_loss": 0.4611632823944092,
|
| 345 |
+
"eval_runtime": 30.4634,
|
| 346 |
+
"eval_samples_per_second": 27.115,
|
| 347 |
+
"eval_steps_per_second": 6.795,
|
| 348 |
+
"eval_wer": 0.3944537413339708,
|
| 349 |
+
"step": 10260
|
| 350 |
+
},
|
| 351 |
+
{
|
| 352 |
+
"epoch": 21.0,
|
| 353 |
+
"grad_norm": 0.029411843046545982,
|
| 354 |
+
"learning_rate": 7.978740157480316e-05,
|
| 355 |
+
"loss": 0.0434,
|
| 356 |
+
"step": 10773
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 21.0,
|
| 360 |
+
"eval_cer": 0.08431599107298311,
|
| 361 |
+
"eval_loss": 0.48810598254203796,
|
| 362 |
+
"eval_runtime": 30.3298,
|
| 363 |
+
"eval_samples_per_second": 27.234,
|
| 364 |
+
"eval_steps_per_second": 6.825,
|
| 365 |
+
"eval_wer": 0.40019125029882857,
|
| 366 |
+
"step": 10773
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 22.0,
|
| 370 |
+
"grad_norm": 0.19160379469394684,
|
| 371 |
+
"learning_rate": 7.877755905511811e-05,
|
| 372 |
+
"loss": 0.0386,
|
| 373 |
+
"step": 11286
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 22.0,
|
| 377 |
+
"eval_cer": 0.08035898004083636,
|
| 378 |
+
"eval_loss": 0.509917676448822,
|
| 379 |
+
"eval_runtime": 31.249,
|
| 380 |
+
"eval_samples_per_second": 26.433,
|
| 381 |
+
"eval_steps_per_second": 6.624,
|
| 382 |
+
"eval_wer": 0.377719340186469,
|
| 383 |
+
"step": 11286
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"epoch": 23.0,
|
| 387 |
+
"grad_norm": 0.18706804513931274,
|
| 388 |
+
"learning_rate": 7.776771653543308e-05,
|
| 389 |
+
"loss": 0.0381,
|
| 390 |
+
"step": 11799
|
| 391 |
+
},
|
| 392 |
+
{
|
| 393 |
+
"epoch": 23.0,
|
| 394 |
+
"eval_cer": 0.08241662577755267,
|
| 395 |
+
"eval_loss": 0.49042966961860657,
|
| 396 |
+
"eval_runtime": 30.3145,
|
| 397 |
+
"eval_samples_per_second": 27.248,
|
| 398 |
+
"eval_steps_per_second": 6.828,
|
| 399 |
+
"eval_wer": 0.38656466650729143,
|
| 400 |
+
"step": 11799
|
| 401 |
+
},
|
| 402 |
+
{
|
| 403 |
+
"epoch": 24.0,
|
| 404 |
+
"grad_norm": 0.048221975564956665,
|
| 405 |
+
"learning_rate": 7.675984251968504e-05,
|
| 406 |
+
"loss": 0.0344,
|
| 407 |
+
"step": 12312
|
| 408 |
+
},
|
| 409 |
+
{
|
| 410 |
+
"epoch": 24.0,
|
| 411 |
+
"eval_cer": 0.08344544864591083,
|
| 412 |
+
"eval_loss": 0.4621882736682892,
|
| 413 |
+
"eval_runtime": 30.4774,
|
| 414 |
+
"eval_samples_per_second": 27.102,
|
| 415 |
+
"eval_steps_per_second": 6.792,
|
| 416 |
+
"eval_wer": 0.4028209419077217,
|
| 417 |
+
"step": 12312
|
| 418 |
+
},
|
| 419 |
+
{
|
| 420 |
+
"epoch": 25.0,
|
| 421 |
+
"grad_norm": 1.851014494895935,
|
| 422 |
+
"learning_rate": 7.575e-05,
|
| 423 |
+
"loss": 0.0302,
|
| 424 |
+
"step": 12825
|
| 425 |
+
},
|
| 426 |
+
{
|
| 427 |
+
"epoch": 25.0,
|
| 428 |
+
"eval_cer": 0.08198926858608081,
|
| 429 |
+
"eval_loss": 0.4985555112361908,
|
| 430 |
+
"eval_runtime": 30.5557,
|
| 431 |
+
"eval_samples_per_second": 27.033,
|
| 432 |
+
"eval_steps_per_second": 6.775,
|
| 433 |
+
"eval_wer": 0.39182404972507767,
|
| 434 |
+
"step": 12825
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 26.0,
|
| 438 |
+
"grad_norm": 0.03737611323595047,
|
| 439 |
+
"learning_rate": 7.474015748031497e-05,
|
| 440 |
+
"loss": 0.0268,
|
| 441 |
+
"step": 13338
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 26.0,
|
| 445 |
+
"eval_cer": 0.08118203833552289,
|
| 446 |
+
"eval_loss": 0.5162001252174377,
|
| 447 |
+
"eval_runtime": 30.3567,
|
| 448 |
+
"eval_samples_per_second": 27.21,
|
| 449 |
+
"eval_steps_per_second": 6.819,
|
| 450 |
+
"eval_wer": 0.3954099928281138,
|
| 451 |
+
"step": 13338
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 27.0,
|
| 455 |
+
"grad_norm": 0.08066482096910477,
|
| 456 |
+
"learning_rate": 7.373031496062992e-05,
|
| 457 |
+
"loss": 0.0272,
|
| 458 |
+
"step": 13851
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 27.0,
|
| 462 |
+
"eval_cer": 0.07914022064293515,
|
| 463 |
+
"eval_loss": 0.4747571647167206,
|
| 464 |
+
"eval_runtime": 30.1451,
|
| 465 |
+
"eval_samples_per_second": 27.401,
|
| 466 |
+
"eval_steps_per_second": 6.867,
|
| 467 |
+
"eval_wer": 0.3773607458761654,
|
| 468 |
+
"step": 13851
|
| 469 |
+
},
|
| 470 |
+
{
|
| 471 |
+
"epoch": 28.0,
|
| 472 |
+
"grad_norm": 1.6533007621765137,
|
| 473 |
+
"learning_rate": 7.27224409448819e-05,
|
| 474 |
+
"loss": 0.0235,
|
| 475 |
+
"step": 14364
|
| 476 |
+
},
|
| 477 |
+
{
|
| 478 |
+
"epoch": 28.0,
|
| 479 |
+
"eval_cer": 0.07853875496604884,
|
| 480 |
+
"eval_loss": 0.47181904315948486,
|
| 481 |
+
"eval_runtime": 30.7257,
|
| 482 |
+
"eval_samples_per_second": 26.883,
|
| 483 |
+
"eval_steps_per_second": 6.737,
|
| 484 |
+
"eval_wer": 0.3822615347836481,
|
| 485 |
+
"step": 14364
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 29.0,
|
| 489 |
+
"grad_norm": 0.020692087709903717,
|
| 490 |
+
"learning_rate": 7.171259842519685e-05,
|
| 491 |
+
"loss": 0.0219,
|
| 492 |
+
"step": 14877
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 29.0,
|
| 496 |
+
"eval_cer": 0.0796942021874357,
|
| 497 |
+
"eval_loss": 0.5317575335502625,
|
| 498 |
+
"eval_runtime": 29.8983,
|
| 499 |
+
"eval_samples_per_second": 27.627,
|
| 500 |
+
"eval_steps_per_second": 6.923,
|
| 501 |
+
"eval_wer": 0.37377480277312936,
|
| 502 |
+
"step": 14877
|
| 503 |
+
},
|
| 504 |
+
{
|
| 505 |
+
"epoch": 30.0,
|
| 506 |
+
"grad_norm": 0.030614351853728294,
|
| 507 |
+
"learning_rate": 7.070275590551182e-05,
|
| 508 |
+
"loss": 0.0205,
|
| 509 |
+
"step": 15390
|
| 510 |
+
},
|
| 511 |
+
{
|
| 512 |
+
"epoch": 30.0,
|
| 513 |
+
"eval_cer": 0.07830133430412003,
|
| 514 |
+
"eval_loss": 0.5195885300636292,
|
| 515 |
+
"eval_runtime": 30.3135,
|
| 516 |
+
"eval_samples_per_second": 27.249,
|
| 517 |
+
"eval_steps_per_second": 6.829,
|
| 518 |
+
"eval_wer": 0.37688262012909396,
|
| 519 |
+
"step": 15390
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 31.0,
|
| 523 |
+
"grad_norm": 1.4625964164733887,
|
| 524 |
+
"learning_rate": 6.969291338582678e-05,
|
| 525 |
+
"loss": 0.0203,
|
| 526 |
+
"step": 15903
|
| 527 |
+
},
|
| 528 |
+
{
|
| 529 |
+
"epoch": 31.0,
|
| 530 |
+
"eval_cer": 0.07879200367210623,
|
| 531 |
+
"eval_loss": 0.5203195214271545,
|
| 532 |
+
"eval_runtime": 29.9013,
|
| 533 |
+
"eval_samples_per_second": 27.624,
|
| 534 |
+
"eval_steps_per_second": 6.923,
|
| 535 |
+
"eval_wer": 0.37413339708343296,
|
| 536 |
+
"step": 15903
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 32.0,
|
| 540 |
+
"grad_norm": 0.2324291467666626,
|
| 541 |
+
"learning_rate": 6.868307086614174e-05,
|
| 542 |
+
"loss": 0.019,
|
| 543 |
+
"step": 16416
|
| 544 |
+
},
|
| 545 |
+
{
|
| 546 |
+
"epoch": 32.0,
|
| 547 |
+
"eval_cer": 0.08089713354120831,
|
| 548 |
+
"eval_loss": 0.5031083822250366,
|
| 549 |
+
"eval_runtime": 30.2,
|
| 550 |
+
"eval_samples_per_second": 27.351,
|
| 551 |
+
"eval_steps_per_second": 6.854,
|
| 552 |
+
"eval_wer": 0.3858474778866842,
|
| 553 |
+
"step": 16416
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 33.0,
|
| 557 |
+
"grad_norm": 0.04213400557637215,
|
| 558 |
+
"learning_rate": 6.76732283464567e-05,
|
| 559 |
+
"loss": 0.0179,
|
| 560 |
+
"step": 16929
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 33.0,
|
| 564 |
+
"eval_cer": 0.08099210180597984,
|
| 565 |
+
"eval_loss": 0.5772231221199036,
|
| 566 |
+
"eval_runtime": 30.1355,
|
| 567 |
+
"eval_samples_per_second": 27.41,
|
| 568 |
+
"eval_steps_per_second": 6.869,
|
| 569 |
+
"eval_wer": 0.37449199139373657,
|
| 570 |
+
"step": 16929
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 34.0,
|
| 574 |
+
"grad_norm": 1.1665908098220825,
|
| 575 |
+
"learning_rate": 6.666338582677165e-05,
|
| 576 |
+
"loss": 0.0175,
|
| 577 |
+
"step": 17442
|
| 578 |
+
},
|
| 579 |
+
{
|
| 580 |
+
"epoch": 34.0,
|
| 581 |
+
"eval_cer": 0.07630700074391808,
|
| 582 |
+
"eval_loss": 0.4905574917793274,
|
| 583 |
+
"eval_runtime": 30.2826,
|
| 584 |
+
"eval_samples_per_second": 27.276,
|
| 585 |
+
"eval_steps_per_second": 6.836,
|
| 586 |
+
"eval_wer": 0.3675591680612001,
|
| 587 |
+
"step": 17442
|
| 588 |
+
},
|
| 589 |
+
{
|
| 590 |
+
"epoch": 35.0,
|
| 591 |
+
"grad_norm": 0.009321126155555248,
|
| 592 |
+
"learning_rate": 6.565354330708662e-05,
|
| 593 |
+
"loss": 0.0166,
|
| 594 |
+
"step": 17955
|
| 595 |
+
},
|
| 596 |
+
{
|
| 597 |
+
"epoch": 35.0,
|
| 598 |
+
"eval_cer": 0.0786020671425632,
|
| 599 |
+
"eval_loss": 0.5370538830757141,
|
| 600 |
+
"eval_runtime": 30.0507,
|
| 601 |
+
"eval_samples_per_second": 27.487,
|
| 602 |
+
"eval_steps_per_second": 6.888,
|
| 603 |
+
"eval_wer": 0.3693521396127181,
|
| 604 |
+
"step": 17955
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 36.0,
|
| 608 |
+
"grad_norm": 0.08575043827295303,
|
| 609 |
+
"learning_rate": 6.464370078740158e-05,
|
| 610 |
+
"loss": 0.0138,
|
| 611 |
+
"step": 18468
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 36.0,
|
| 615 |
+
"eval_cer": 0.07882365976036342,
|
| 616 |
+
"eval_loss": 0.5748464465141296,
|
| 617 |
+
"eval_runtime": 30.2194,
|
| 618 |
+
"eval_samples_per_second": 27.333,
|
| 619 |
+
"eval_steps_per_second": 6.85,
|
| 620 |
+
"eval_wer": 0.37437245995696866,
|
| 621 |
+
"step": 18468
|
| 622 |
+
},
|
| 623 |
+
{
|
| 624 |
+
"epoch": 37.0,
|
| 625 |
+
"grad_norm": 0.006642826832830906,
|
| 626 |
+
"learning_rate": 6.363582677165355e-05,
|
| 627 |
+
"loss": 0.0134,
|
| 628 |
+
"step": 18981
|
| 629 |
+
},
|
| 630 |
+
{
|
| 631 |
+
"epoch": 37.0,
|
| 632 |
+
"eval_cer": 0.07782649298026242,
|
| 633 |
+
"eval_loss": 0.5342956781387329,
|
| 634 |
+
"eval_runtime": 30.2477,
|
| 635 |
+
"eval_samples_per_second": 27.308,
|
| 636 |
+
"eval_steps_per_second": 6.843,
|
| 637 |
+
"eval_wer": 0.36971073392302173,
|
| 638 |
+
"step": 18981
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 38.0,
|
| 642 |
+
"grad_norm": 0.032926104962825775,
|
| 643 |
+
"learning_rate": 6.26259842519685e-05,
|
| 644 |
+
"loss": 0.0135,
|
| 645 |
+
"step": 19494
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 38.0,
|
| 649 |
+
"eval_cer": 0.08035898004083636,
|
| 650 |
+
"eval_loss": 0.5406966805458069,
|
| 651 |
+
"eval_runtime": 30.4073,
|
| 652 |
+
"eval_samples_per_second": 27.164,
|
| 653 |
+
"eval_steps_per_second": 6.808,
|
| 654 |
+
"eval_wer": 0.3839349748983983,
|
| 655 |
+
"step": 19494
|
| 656 |
+
},
|
| 657 |
+
{
|
| 658 |
+
"epoch": 39.0,
|
| 659 |
+
"grad_norm": 0.4430481791496277,
|
| 660 |
+
"learning_rate": 6.161614173228348e-05,
|
| 661 |
+
"loss": 0.0123,
|
| 662 |
+
"step": 20007
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 39.0,
|
| 666 |
+
"eval_cer": 0.07668687380300417,
|
| 667 |
+
"eval_loss": 0.5342694520950317,
|
| 668 |
+
"eval_runtime": 29.877,
|
| 669 |
+
"eval_samples_per_second": 27.647,
|
| 670 |
+
"eval_steps_per_second": 6.928,
|
| 671 |
+
"eval_wer": 0.36612479081998567,
|
| 672 |
+
"step": 20007
|
| 673 |
+
},
|
| 674 |
+
{
|
| 675 |
+
"epoch": 40.0,
|
| 676 |
+
"grad_norm": 0.005585499573498964,
|
| 677 |
+
"learning_rate": 6.0606299212598426e-05,
|
| 678 |
+
"loss": 0.0124,
|
| 679 |
+
"step": 20520
|
| 680 |
+
},
|
| 681 |
+
{
|
| 682 |
+
"epoch": 40.0,
|
| 683 |
+
"eval_cer": 0.08170436379176625,
|
| 684 |
+
"eval_loss": 0.5632651448249817,
|
| 685 |
+
"eval_runtime": 30.3597,
|
| 686 |
+
"eval_samples_per_second": 27.207,
|
| 687 |
+
"eval_steps_per_second": 6.818,
|
| 688 |
+
"eval_wer": 0.3801099689218264,
|
| 689 |
+
"step": 20520
|
| 690 |
+
},
|
| 691 |
+
{
|
| 692 |
+
"epoch": 41.0,
|
| 693 |
+
"grad_norm": 0.006700966972857714,
|
| 694 |
+
"learning_rate": 5.9596456692913396e-05,
|
| 695 |
+
"loss": 0.0131,
|
| 696 |
+
"step": 21033
|
| 697 |
+
},
|
| 698 |
+
{
|
| 699 |
+
"epoch": 41.0,
|
| 700 |
+
"eval_cer": 0.07735165165640481,
|
| 701 |
+
"eval_loss": 0.5580958724021912,
|
| 702 |
+
"eval_runtime": 30.4235,
|
| 703 |
+
"eval_samples_per_second": 27.15,
|
| 704 |
+
"eval_steps_per_second": 6.804,
|
| 705 |
+
"eval_wer": 0.36325603633755676,
|
| 706 |
+
"step": 21033
|
| 707 |
+
},
|
| 708 |
+
{
|
| 709 |
+
"epoch": 42.0,
|
| 710 |
+
"grad_norm": 0.0002161410520784557,
|
| 711 |
+
"learning_rate": 5.858661417322835e-05,
|
| 712 |
+
"loss": 0.0094,
|
| 713 |
+
"step": 21546
|
| 714 |
+
},
|
| 715 |
+
{
|
| 716 |
+
"epoch": 42.0,
|
| 717 |
+
"eval_cer": 0.07888697193687776,
|
| 718 |
+
"eval_loss": 0.5862228870391846,
|
| 719 |
+
"eval_runtime": 30.2348,
|
| 720 |
+
"eval_samples_per_second": 27.319,
|
| 721 |
+
"eval_steps_per_second": 6.846,
|
| 722 |
+
"eval_wer": 0.3683958881185752,
|
| 723 |
+
"step": 21546
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 43.0,
|
| 727 |
+
"grad_norm": 0.005379769951105118,
|
| 728 |
+
"learning_rate": 5.75767716535433e-05,
|
| 729 |
+
"loss": 0.0101,
|
| 730 |
+
"step": 22059
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 43.0,
|
| 734 |
+
"eval_cer": 0.07608540812611786,
|
| 735 |
+
"eval_loss": 0.5479421615600586,
|
| 736 |
+
"eval_runtime": 30.4733,
|
| 737 |
+
"eval_samples_per_second": 27.106,
|
| 738 |
+
"eval_steps_per_second": 6.793,
|
| 739 |
+
"eval_wer": 0.36457088214200334,
|
| 740 |
+
"step": 22059
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 44.0,
|
| 744 |
+
"grad_norm": 0.008168138563632965,
|
| 745 |
+
"learning_rate": 5.656692913385827e-05,
|
| 746 |
+
"loss": 0.0094,
|
| 747 |
+
"step": 22572
|
| 748 |
+
},
|
| 749 |
+
{
|
| 750 |
+
"epoch": 44.0,
|
| 751 |
+
"eval_cer": 0.07605375203786068,
|
| 752 |
+
"eval_loss": 0.5737643241882324,
|
| 753 |
+
"eval_runtime": 30.1141,
|
| 754 |
+
"eval_samples_per_second": 27.429,
|
| 755 |
+
"eval_steps_per_second": 6.874,
|
| 756 |
+
"eval_wer": 0.3620607219698781,
|
| 757 |
+
"step": 22572
|
| 758 |
+
},
|
| 759 |
+
{
|
| 760 |
+
"epoch": 45.0,
|
| 761 |
+
"grad_norm": 0.001390150049701333,
|
| 762 |
+
"learning_rate": 5.555708661417323e-05,
|
| 763 |
+
"loss": 0.0078,
|
| 764 |
+
"step": 23085
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 45.0,
|
| 768 |
+
"eval_cer": 0.0777473527596195,
|
| 769 |
+
"eval_loss": 0.5284022688865662,
|
| 770 |
+
"eval_runtime": 30.2759,
|
| 771 |
+
"eval_samples_per_second": 27.282,
|
| 772 |
+
"eval_steps_per_second": 6.837,
|
| 773 |
+
"eval_wer": 0.37819746593354053,
|
| 774 |
+
"step": 23085
|
| 775 |
+
},
|
| 776 |
+
{
|
| 777 |
+
"epoch": 46.0,
|
| 778 |
+
"grad_norm": NaN,
|
| 779 |
+
"learning_rate": 5.4551181102362206e-05,
|
| 780 |
+
"loss": 0.0074,
|
| 781 |
+
"step": 23598
|
| 782 |
+
},
|
| 783 |
+
{
|
| 784 |
+
"epoch": 46.0,
|
| 785 |
+
"eval_cer": 0.07898194020164928,
|
| 786 |
+
"eval_loss": 0.627711832523346,
|
| 787 |
+
"eval_runtime": 30.4473,
|
| 788 |
+
"eval_samples_per_second": 27.129,
|
| 789 |
+
"eval_steps_per_second": 6.799,
|
| 790 |
+
"eval_wer": 0.3724599569686828,
|
| 791 |
+
"step": 23598
|
| 792 |
+
},
|
| 793 |
+
{
|
| 794 |
+
"epoch": 47.0,
|
| 795 |
+
"grad_norm": 0.0076849376782774925,
|
| 796 |
+
"learning_rate": 5.354133858267717e-05,
|
| 797 |
+
"loss": 0.01,
|
| 798 |
+
"step": 24111
|
| 799 |
+
},
|
| 800 |
+
{
|
| 801 |
+
"epoch": 47.0,
|
| 802 |
+
"eval_cer": 0.0765127653175897,
|
| 803 |
+
"eval_loss": 0.5825507640838623,
|
| 804 |
+
"eval_runtime": 30.4404,
|
| 805 |
+
"eval_samples_per_second": 27.135,
|
| 806 |
+
"eval_steps_per_second": 6.8,
|
| 807 |
+
"eval_wer": 0.3686349509921109,
|
| 808 |
+
"step": 24111
|
| 809 |
+
},
|
| 810 |
+
{
|
| 811 |
+
"epoch": 48.0,
|
| 812 |
+
"grad_norm": 0.015365912578999996,
|
| 813 |
+
"learning_rate": 5.2531496062992125e-05,
|
| 814 |
+
"loss": 0.0088,
|
| 815 |
+
"step": 24624
|
| 816 |
+
},
|
| 817 |
+
{
|
| 818 |
+
"epoch": 48.0,
|
| 819 |
+
"eval_cer": 0.07611706421437503,
|
| 820 |
+
"eval_loss": 0.5601234436035156,
|
| 821 |
+
"eval_runtime": 30.4389,
|
| 822 |
+
"eval_samples_per_second": 27.136,
|
| 823 |
+
"eval_steps_per_second": 6.801,
|
| 824 |
+
"eval_wer": 0.36600525938321776,
|
| 825 |
+
"step": 24624
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 49.0,
|
| 829 |
+
"grad_norm": 0.0013850124087184668,
|
| 830 |
+
"learning_rate": 5.152165354330709e-05,
|
| 831 |
+
"loss": 0.0083,
|
| 832 |
+
"step": 25137
|
| 833 |
+
},
|
| 834 |
+
{
|
| 835 |
+
"epoch": 49.0,
|
| 836 |
+
"eval_cer": 0.07686098228841862,
|
| 837 |
+
"eval_loss": 0.5409994721412659,
|
| 838 |
+
"eval_runtime": 30.7659,
|
| 839 |
+
"eval_samples_per_second": 26.848,
|
| 840 |
+
"eval_steps_per_second": 6.728,
|
| 841 |
+
"eval_wer": 0.3606263447286636,
|
| 842 |
+
"step": 25137
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 50.0,
|
| 846 |
+
"grad_norm": 0.0006066646310500801,
|
| 847 |
+
"learning_rate": 5.0513779527559055e-05,
|
| 848 |
+
"loss": 0.0074,
|
| 849 |
+
"step": 25650
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 50.0,
|
| 853 |
+
"eval_cer": 0.07804808559806264,
|
| 854 |
+
"eval_loss": 0.559197723865509,
|
| 855 |
+
"eval_runtime": 30.2523,
|
| 856 |
+
"eval_samples_per_second": 27.304,
|
| 857 |
+
"eval_steps_per_second": 6.842,
|
| 858 |
+
"eval_wer": 0.3613435333492709,
|
| 859 |
+
"step": 25650
|
| 860 |
+
},
|
| 861 |
+
{
|
| 862 |
+
"epoch": 51.0,
|
| 863 |
+
"grad_norm": 0.04679730907082558,
|
| 864 |
+
"learning_rate": 4.950393700787402e-05,
|
| 865 |
+
"loss": 0.007,
|
| 866 |
+
"step": 26163
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 51.0,
|
| 870 |
+
"eval_cer": 0.07793728928916253,
|
| 871 |
+
"eval_loss": 0.5890862941741943,
|
| 872 |
+
"eval_runtime": 30.6371,
|
| 873 |
+
"eval_samples_per_second": 26.961,
|
| 874 |
+
"eval_steps_per_second": 6.757,
|
| 875 |
+
"eval_wer": 0.3689935453024145,
|
| 876 |
+
"step": 26163
|
| 877 |
+
},
|
| 878 |
+
{
|
| 879 |
+
"epoch": 52.0,
|
| 880 |
+
"grad_norm": 0.012076578103005886,
|
| 881 |
+
"learning_rate": 4.849409448818898e-05,
|
| 882 |
+
"loss": 0.0067,
|
| 883 |
+
"step": 26676
|
| 884 |
+
},
|
| 885 |
+
{
|
| 886 |
+
"epoch": 52.0,
|
| 887 |
+
"eval_cer": 0.07788980515677678,
|
| 888 |
+
"eval_loss": 0.5806664228439331,
|
| 889 |
+
"eval_runtime": 30.6637,
|
| 890 |
+
"eval_samples_per_second": 26.937,
|
| 891 |
+
"eval_steps_per_second": 6.751,
|
| 892 |
+
"eval_wer": 0.3662443222567535,
|
| 893 |
+
"step": 26676
|
| 894 |
+
},
|
| 895 |
+
{
|
| 896 |
+
"epoch": 53.0,
|
| 897 |
+
"grad_norm": 0.0049696676433086395,
|
| 898 |
+
"learning_rate": 4.748425196850394e-05,
|
| 899 |
+
"loss": 0.0067,
|
| 900 |
+
"step": 27189
|
| 901 |
+
},
|
| 902 |
+
{
|
| 903 |
+
"epoch": 53.0,
|
| 904 |
+
"eval_cer": 0.07728833947989047,
|
| 905 |
+
"eval_loss": 0.5851139426231384,
|
| 906 |
+
"eval_runtime": 30.4226,
|
| 907 |
+
"eval_samples_per_second": 27.151,
|
| 908 |
+
"eval_steps_per_second": 6.804,
|
| 909 |
+
"eval_wer": 0.363973224958164,
|
| 910 |
+
"step": 27189
|
| 911 |
+
},
|
| 912 |
+
{
|
| 913 |
+
"epoch": 54.0,
|
| 914 |
+
"grad_norm": 0.0018264329992234707,
|
| 915 |
+
"learning_rate": 4.64744094488189e-05,
|
| 916 |
+
"loss": 0.0065,
|
| 917 |
+
"step": 27702
|
| 918 |
+
},
|
| 919 |
+
{
|
| 920 |
+
"epoch": 54.0,
|
| 921 |
+
"eval_cer": 0.07667104575887558,
|
| 922 |
+
"eval_loss": 0.5988566875457764,
|
| 923 |
+
"eval_runtime": 29.9692,
|
| 924 |
+
"eval_samples_per_second": 27.562,
|
| 925 |
+
"eval_steps_per_second": 6.907,
|
| 926 |
+
"eval_wer": 0.36672244800382503,
|
| 927 |
+
"step": 27702
|
| 928 |
+
},
|
| 929 |
+
{
|
| 930 |
+
"epoch": 55.0,
|
| 931 |
+
"grad_norm": 5.14693021774292,
|
| 932 |
+
"learning_rate": 4.5464566929133864e-05,
|
| 933 |
+
"loss": 0.005,
|
| 934 |
+
"step": 28215
|
| 935 |
+
},
|
| 936 |
+
{
|
| 937 |
+
"epoch": 55.0,
|
| 938 |
+
"eval_cer": 0.07852292692192026,
|
| 939 |
+
"eval_loss": 0.5746394395828247,
|
| 940 |
+
"eval_runtime": 30.2216,
|
| 941 |
+
"eval_samples_per_second": 27.331,
|
| 942 |
+
"eval_steps_per_second": 6.849,
|
| 943 |
+
"eval_wer": 0.37568730576141524,
|
| 944 |
+
"step": 28215
|
| 945 |
+
}
|
| 946 |
+
],
|
| 947 |
+
"logging_steps": 500,
|
| 948 |
+
"max_steps": 51300,
|
| 949 |
+
"num_input_tokens_seen": 0,
|
| 950 |
+
"num_train_epochs": 100,
|
| 951 |
+
"save_steps": 500,
|
| 952 |
+
"stateful_callbacks": {
|
| 953 |
+
"EarlyStoppingCallback": {
|
| 954 |
+
"args": {
|
| 955 |
+
"early_stopping_patience": 20,
|
| 956 |
+
"early_stopping_threshold": 0.001
|
| 957 |
+
},
|
| 958 |
+
"attributes": {
|
| 959 |
+
"early_stopping_patience_counter": 6
|
| 960 |
+
}
|
| 961 |
+
},
|
| 962 |
+
"TrainerControl": {
|
| 963 |
+
"args": {
|
| 964 |
+
"should_epoch_stop": false,
|
| 965 |
+
"should_evaluate": false,
|
| 966 |
+
"should_log": false,
|
| 967 |
+
"should_save": true,
|
| 968 |
+
"should_training_stop": false
|
| 969 |
+
},
|
| 970 |
+
"attributes": {}
|
| 971 |
+
}
|
| 972 |
+
},
|
| 973 |
+
"total_flos": 6.001178563865598e+19,
|
| 974 |
+
"train_batch_size": 8,
|
| 975 |
+
"trial_name": null,
|
| 976 |
+
"trial_params": null
|
| 977 |
+
}
|
checkpoint-28215/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2b71a1c487e8560b2d69fe2832962ae0eb6d2d0e7304ab8707b68c92ffe58d66
|
| 3 |
+
size 5304
|