Training in progress, epoch 9, checkpoint

checkpoint-4617/config.json

@@ -0,0 +1,82 @@
+{
+  "_name_or_path": "facebook/w2v-bert-2.0",
+  "activation_dropout": 0.0,
+  "adapter_act": "relu",
+  "adapter_kernel_size": 3,
+  "adapter_stride": 2,
+  "add_adapter": true,
+  "apply_spec_augment": false,
+  "architectures": [
+    "Wav2Vec2BertForCTC"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 1,
+  "classifier_proj_size": 768,
+  "codevector_dim": 768,
+  "conformer_conv_dropout": 0.1,
+  "contrastive_logits_temperature": 0.1,
+  "conv_depthwise_kernel_size": 31,
+  "ctc_loss_reduction": "mean",
+  "ctc_zero_infinity": true,
+  "diversity_loss_weight": 0.1,
+  "eos_token_id": 2,
+  "feat_proj_dropout": 0.0,
+  "feat_quantizer_dropout": 0.0,
+  "feature_projection_input_dim": 160,
+  "final_dropout": 0.1,
+  "hidden_act": "swish",
+  "hidden_dropout": 0.0,
+  "hidden_size": 1024,
+  "initializer_range": 0.02,
+  "intermediate_size": 4096,
+  "layer_norm_eps": 1e-05,
+  "layerdrop": 0.0,
+  "left_max_position_embeddings": 64,
+  "mask_feature_length": 10,
+  "mask_feature_min_masks": 0,
+  "mask_feature_prob": 0.0,
+  "mask_time_length": 10,
+  "mask_time_min_masks": 2,
+  "mask_time_prob": 0.0,
+  "max_source_positions": 5000,
+  "model_type": "wav2vec2-bert",
+  "num_adapter_layers": 1,
+  "num_attention_heads": 16,
+  "num_codevector_groups": 2,
+  "num_codevectors_per_group": 320,
+  "num_hidden_layers": 24,
+  "num_negatives": 100,
+  "output_hidden_size": 1024,
+  "pad_token_id": 45,
+  "position_embeddings_type": "relative_key",
+  "proj_codevector_dim": 768,
+  "right_max_position_embeddings": 8,
+  "rotary_embedding_base": 10000,
+  "tdnn_dilation": [
+    1,
+    2,
+    3,
+    1,
+    1
+  ],
+  "tdnn_dim": [
+    512,
+    512,
+    512,
+    512,
+    1500
+  ],
+  "tdnn_kernel": [
+    5,
+    3,
+    3,
+    1,
+    1
+  ],
+  "torch_dtype": "float32",
+  "transformers_version": "4.46.0",
+  "use_intermediate_ffn_before_adapter": false,
+  "use_weighted_layer_sum": false,
+  "vocab_size": 48,
+  "xvector_output_dim": 512
+}

checkpoint-4617/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:479e5625a937272d83c756dd3d94bd31380cc3024c16ff61e326e0522a5f34d9
+size 2423011360

checkpoint-4617/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b520173dc6d725c08275284d4650ad66c7b065185f9cf6a0c556dc41ec4d4e45
+size 4846483562

checkpoint-4617/preprocessor_config.json

@@ -0,0 +1,11 @@
+{
+  "feature_extractor_type": "SeamlessM4TFeatureExtractor",
+  "feature_size": 80,
+  "num_mel_bins": 80,
+  "padding_side": "right",
+  "padding_value": 1,
+  "processor_class": "Wav2Vec2BertProcessor",
+  "return_attention_mask": true,
+  "sampling_rate": 16000,
+  "stride": 2
+}

checkpoint-4617/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bd7b4704ac33a98aa7ba8c5171f638e8c93b20427bfd35500311ea6eab0a2b8e
+size 14244

checkpoint-4617/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:792e8f6c60f1efbff039abeff809b24bc7e33e42a3b814ff5742085d3d0ec615
+size 1064

checkpoint-4617/trainer_state.json

@@ -0,0 +1,195 @@
+{
+  "best_metric": 0.4000717188620607,
+  "best_model_checkpoint": "w2v-bert-2.0-CV_Fleurs-lg-10hrs-v4/checkpoint-3591",
+  "epoch": 9.0,
+  "eval_steps": 500,
+  "global_step": 4617,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 1.0,
+      "grad_norm": 30.792112350463867,
+      "learning_rate": 9.998031496062993e-05,
+      "loss": 1.4106,
+      "step": 513
+    },
+    {
+      "epoch": 1.0,
+      "eval_cer": 0.1264185884550246,
+      "eval_loss": 0.5293365716934204,
+      "eval_runtime": 31.0205,
+      "eval_samples_per_second": 26.628,
+      "eval_steps_per_second": 6.673,
+      "eval_wer": 0.5665790102797036,
+      "step": 513
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 10.16055679321289,
+      "learning_rate": 9.897047244094488e-05,
+      "loss": 0.4775,
+      "step": 1026
+    },
+    {
+      "epoch": 2.0,
+      "eval_cer": 0.11182513176846737,
+      "eval_loss": 0.4882170259952545,
+      "eval_runtime": 30.4398,
+      "eval_samples_per_second": 27.135,
+      "eval_steps_per_second": 6.8,
+      "eval_wer": 0.5251016017212526,
+      "step": 1026
+    },
+    {
+      "epoch": 3.0,
+      "grad_norm": 24.60751724243164,
+      "learning_rate": 9.796062992125985e-05,
+      "loss": 0.3696,
+      "step": 1539
+    },
+    {
+      "epoch": 3.0,
+      "eval_cer": 0.09981164627486981,
+      "eval_loss": 0.397329181432724,
+      "eval_runtime": 30.5019,
+      "eval_samples_per_second": 27.08,
+      "eval_steps_per_second": 6.786,
+      "eval_wer": 0.4564905570164953,
+      "step": 1539
+    },
+    {
+      "epoch": 4.0,
+      "grad_norm": 8.973501205444336,
+      "learning_rate": 9.695078740157481e-05,
+      "loss": 0.3007,
+      "step": 2052
+    },
+    {
+      "epoch": 4.0,
+      "eval_cer": 0.09322717991737761,
+      "eval_loss": 0.3895888030529022,
+      "eval_runtime": 29.8209,
+      "eval_samples_per_second": 27.699,
+      "eval_steps_per_second": 6.941,
+      "eval_wer": 0.4311498924217069,
+      "step": 2052
+    },
+    {
+      "epoch": 5.0,
+      "grad_norm": 3.461097240447998,
+      "learning_rate": 9.594094488188978e-05,
+      "loss": 0.2564,
+      "step": 2565
+    },
+    {
+      "epoch": 5.0,
+      "eval_cer": 0.09659855331676664,
+      "eval_loss": 0.3743041455745697,
+      "eval_runtime": 29.9644,
+      "eval_samples_per_second": 27.566,
+      "eval_steps_per_second": 6.908,
+      "eval_wer": 0.4372459956968683,
+      "step": 2565
+    },
+    {
+      "epoch": 6.0,
+      "grad_norm": 9.60593032836914,
+      "learning_rate": 9.493110236220472e-05,
+      "loss": 0.2273,
+      "step": 3078
+    },
+    {
+      "epoch": 6.0,
+      "eval_cer": 0.0879880973108153,
+      "eval_loss": 0.3549309968948364,
+      "eval_runtime": 30.2119,
+      "eval_samples_per_second": 27.34,
+      "eval_steps_per_second": 6.852,
+      "eval_wer": 0.41836002868754485,
+      "step": 3078
+    },
+    {
+      "epoch": 7.0,
+      "grad_norm": 2.7023348808288574,
+      "learning_rate": 9.392125984251969e-05,
+      "loss": 0.203,
+      "step": 3591
+    },
+    {
+      "epoch": 7.0,
+      "eval_cer": 0.08468003608794061,
+      "eval_loss": 0.3835783302783966,
+      "eval_runtime": 29.6315,
+      "eval_samples_per_second": 27.876,
+      "eval_steps_per_second": 6.986,
+      "eval_wer": 0.4000717188620607,
+      "step": 3591
+    },
+    {
+      "epoch": 8.0,
+      "grad_norm": 2.196669340133667,
+      "learning_rate": 9.291141732283465e-05,
+      "loss": 0.1775,
+      "step": 4104
+    },
+    {
+      "epoch": 8.0,
+      "eval_cer": 0.08550309438262714,
+      "eval_loss": 0.3693934977054596,
+      "eval_runtime": 30.2095,
+      "eval_samples_per_second": 27.342,
+      "eval_steps_per_second": 6.852,
+      "eval_wer": 0.4132201769065264,
+      "step": 4104
+    },
+    {
+      "epoch": 9.0,
+      "grad_norm": 3.045036792755127,
+      "learning_rate": 9.190354330708662e-05,
+      "loss": 0.159,
+      "step": 4617
+    },
+    {
+      "epoch": 9.0,
+      "eval_cer": 0.0889694360467877,
+      "eval_loss": 0.3487217128276825,
+      "eval_runtime": 29.9136,
+      "eval_samples_per_second": 27.613,
+      "eval_steps_per_second": 6.92,
+      "eval_wer": 0.4136983026535979,
+      "step": 4617
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 51300,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 100,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "EarlyStoppingCallback": {
+      "args": {
+        "early_stopping_patience": 20,
+        "early_stopping_threshold": 0.001
+      },
+      "attributes": {
+        "early_stopping_patience_counter": 2
+      }
+    },
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 9.820287690574264e+18,
+  "train_batch_size": 8,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-4617/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2b71a1c487e8560b2d69fe2832962ae0eb6d2d0e7304ab8707b68c92ffe58d66
+size 5304