diff --git a/README.md b/README.md new file mode 100644 index 0000000000000000000000000000000000000000..e8700f1fb8efe368c231e679684d572a99e6ac74 --- /dev/null +++ b/README.md @@ -0,0 +1,79 @@ +--- +license: apache-2.0 +tags: +- generated_from_trainer +datasets: +- audiofolder +metrics: +- accuracy +- f1 +- recall +- precision +model-index: +- name: wav2vec2-base-is_vinyl_scratched_or_not + results: [] +--- + + + +# wav2vec2-base-is_vinyl_scratched_or_not + +This model is a fine-tuned version of [facebook/wav2vec2-base](https://huggingface.co/facebook/wav2vec2-base) on the audiofolder dataset. +It achieves the following results on the evaluation set: +- Loss: 0.1039 +- Accuracy: 0.9752 +- F1: 0.9638 +- Recall: 0.9576 +- Precision: 0.9700 + +## Model description + +More information needed + +## Intended uses & limitations + +More information needed + +## Training and evaluation data + +More information needed + +## Training procedure + +### Training hyperparameters + +The following hyperparameters were used during training: +- learning_rate: 3e-05 +- train_batch_size: 32 +- eval_batch_size: 32 +- seed: 42 +- gradient_accumulation_steps: 4 +- total_train_batch_size: 128 +- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 +- lr_scheduler_type: linear +- lr_scheduler_warmup_ratio: 0.1 +- num_epochs: 10 + +### Training results + +| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 | Recall | Precision | +|:-------------:|:-----:|:----:|:---------------:|:--------:|:------:|:------:|:---------:| +| 0.6671 | 0.98 | 21 | 0.6235 | 0.6560 | 0.0 | 0.0 | 0.0 | +| 0.4954 | 1.98 | 42 | 0.2824 | 0.9417 | 0.9095 | 0.8517 | 0.9757 | +| 0.2406 | 2.98 | 63 | 0.1755 | 0.9563 | 0.9336 | 0.8941 | 0.9769 | +| 0.169 | 3.98 | 84 | 0.1545 | 0.9592 | 0.9386 | 0.9068 | 0.9727 | +| 0.1287 | 4.98 | 105 | 0.1249 | 0.9606 | 0.9407 | 0.9068 | 0.9772 | +| 0.1102 | 5.98 | 126 | 0.1159 | 0.9723 | 0.9595 | 0.9534 | 0.9657 | +| 0.0923 | 6.98 | 147 | 0.1073 | 0.9665 | 0.9516 | 0.9576 | 0.9456 | +| 0.0877 | 7.98 | 168 | 0.1039 | 0.9752 | 0.9638 | 0.9576 | 0.9700 | +| 0.0807 | 8.98 | 189 | 0.1088 | 0.9679 | 0.9536 | 0.9576 | 0.9496 | +| 0.0744 | 9.98 | 210 | 0.1041 | 0.9752 | 0.9638 | 0.9576 | 0.9700 | + + +### Framework versions + +- Transformers 4.26.0 +- Pytorch 1.12.1 +- Datasets 2.8.0 +- Tokenizers 0.12.1 diff --git a/checkpoint-105/config.json b/checkpoint-105/config.json new file mode 100644 index 0000000000000000000000000000000000000000..5895b668646a423372f040c92bb60d42a5449727 --- /dev/null +++ b/checkpoint-105/config.json @@ -0,0 +1,125 @@ +{ + "_name_or_path": "facebook/wav2vec2-base", + "activation_dropout": 0.0, + "adapter_kernel_size": 3, + "adapter_stride": 2, + "add_adapter": false, + "apply_spec_augment": true, + "architectures": [ + "Wav2Vec2ForSequenceClassification" + ], + "attention_dropout": 0.1, + "bos_token_id": 1, + "classifier_proj_size": 256, + "codevector_dim": 256, + "contrastive_logits_temperature": 0.1, + "conv_bias": false, + "conv_dim": [ + 512, + 512, + 512, + 512, + 512, + 512, + 512 + ], + "conv_kernel": [ + 10, + 3, + 3, + 3, + 3, + 2, + 2 + ], + "conv_stride": [ + 5, + 2, + 2, + 2, + 2, + 2, + 2 + ], + "ctc_loss_reduction": "sum", + "ctc_zero_infinity": false, + "diversity_loss_weight": 0.1, + "do_stable_layer_norm": false, + "eos_token_id": 2, + "feat_extract_activation": "gelu", + "feat_extract_norm": "group", + "feat_proj_dropout": 0.1, + "feat_quantizer_dropout": 0.0, + "final_dropout": 0.0, + "freeze_feat_extract_train": true, + "hidden_act": "gelu", + "hidden_dropout": 0.1, + "hidden_size": 768, + "id2label": { + "0": 0, + "1": 1 + }, + "initializer_range": 0.02, + "intermediate_size": 3072, + "label2id": { + "0": "0", + "1": "1" + }, + "layer_norm_eps": 1e-05, + "layerdrop": 0.0, + "mask_channel_length": 10, + "mask_channel_min_space": 1, + "mask_channel_other": 0.0, + "mask_channel_prob": 0.0, + "mask_channel_selection": "static", + "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_min_space": 1, + "mask_time_other": 0.0, + "mask_time_prob": 0.05, + "mask_time_selection": "static", + "model_type": "wav2vec2", + "no_mask_channel_overlap": false, + "no_mask_time_overlap": false, + "num_adapter_layers": 3, + "num_attention_heads": 12, + "num_codevector_groups": 2, + "num_codevectors_per_group": 320, + "num_conv_pos_embedding_groups": 16, + "num_conv_pos_embeddings": 128, + "num_feat_extract_layers": 7, + "num_hidden_layers": 12, + "num_negatives": 100, + "output_hidden_size": 768, + "pad_token_id": 0, + "proj_codevector_dim": 256, + "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": "4.26.0", + "use_weighted_layer_sum": false, + "vocab_size": 32, + "xvector_output_dim": 512 +} diff --git a/checkpoint-105/optimizer.pt b/checkpoint-105/optimizer.pt new file mode 100644 index 0000000000000000000000000000000000000000..93a30455b108ed2fb786acc98dfd662a954f7051 --- /dev/null +++ b/checkpoint-105/optimizer.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4dc615861c1c97db60cc54716d8f22865204d98aeaa117729b5e1e2237c70eeb +size 756671269 diff --git a/checkpoint-105/preprocessor_config.json b/checkpoint-105/preprocessor_config.json new file mode 100644 index 0000000000000000000000000000000000000000..a0b7227fc1d916e469b14f6c154ad6dfea1e6891 --- /dev/null +++ b/checkpoint-105/preprocessor_config.json @@ -0,0 +1,9 @@ +{ + "do_normalize": true, + "feature_extractor_type": "Wav2Vec2FeatureExtractor", + "feature_size": 1, + "padding_side": "right", + "padding_value": 0.0, + "return_attention_mask": false, + "sampling_rate": 16000 +} diff --git a/checkpoint-105/pytorch_model.bin b/checkpoint-105/pytorch_model.bin new file mode 100644 index 0000000000000000000000000000000000000000..fda0bf0171fff27b0563d62885d91f98d5c295a0 --- /dev/null +++ b/checkpoint-105/pytorch_model.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:03e5b21cd48c3f71dc8de5f13d62125b5f37b1007e5375a55acd0ca79b5d8013 +size 378345555 diff --git a/checkpoint-105/rng_state.pth b/checkpoint-105/rng_state.pth new file mode 100644 index 0000000000000000000000000000000000000000..ffbe827177d1159a04e5c341961c8b2736b070fd --- /dev/null +++ b/checkpoint-105/rng_state.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d8a2b066fe203264a80066a1515be8c6e3ad6a487e5449a3884608b70bcd6043 +size 13611 diff --git a/checkpoint-105/scheduler.pt b/checkpoint-105/scheduler.pt new file mode 100644 index 0000000000000000000000000000000000000000..77a290d1fb6c7f58e13e57928e7531c4d6b8eec9 --- /dev/null +++ b/checkpoint-105/scheduler.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:802834eb6df0efb5ed147e668771cb8da19ac61fb464214f1fdc4375abceca4e +size 623 diff --git a/checkpoint-105/trainer_state.json b/checkpoint-105/trainer_state.json new file mode 100644 index 0000000000000000000000000000000000000000..cc232a569aca3d0b637d8aadbba27b64608b8109 --- /dev/null +++ b/checkpoint-105/trainer_state.json @@ -0,0 +1,112 @@ +{ + "best_metric": 0.9406593406593408, + "best_model_checkpoint": "wav2vec2-base-is_vinyl_scratched_or_not/checkpoint-105", + "epoch": 4.976744186046512, + "global_step": 105, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.05, + "learning_rate": 1.4285714285714286e-06, + "loss": 0.6901, + "step": 1 + }, + { + "epoch": 0.98, + "learning_rate": 3e-05, + "loss": 0.6671, + "step": 21 + }, + { + "epoch": 0.98, + "eval_F1": 0.0, + "eval_Precision": 0.0, + "eval_Recall": 0.0, + "eval_accuracy": 0.6559766763848397, + "eval_loss": 0.6235492825508118, + "eval_runtime": 1064.9816, + "eval_samples_per_second": 0.644, + "eval_steps_per_second": 0.021, + "step": 21 + }, + { + "epoch": 1.98, + "learning_rate": 2.6666666666666667e-05, + "loss": 0.4954, + "step": 42 + }, + { + "epoch": 1.98, + "eval_F1": 0.9095022624434389, + "eval_Precision": 0.9757281553398058, + "eval_Recall": 0.8516949152542372, + "eval_accuracy": 0.9416909620991254, + "eval_loss": 0.28244873881340027, + "eval_runtime": 1042.1602, + "eval_samples_per_second": 0.658, + "eval_steps_per_second": 0.021, + "step": 42 + }, + { + "epoch": 2.98, + "learning_rate": 2.3333333333333336e-05, + "loss": 0.2406, + "step": 63 + }, + { + "epoch": 2.98, + "eval_F1": 0.9336283185840708, + "eval_Precision": 0.9768518518518519, + "eval_Recall": 0.8940677966101694, + "eval_accuracy": 0.956268221574344, + "eval_loss": 0.17554502189159393, + "eval_runtime": 1040.1013, + "eval_samples_per_second": 0.66, + "eval_steps_per_second": 0.021, + "step": 63 + }, + { + "epoch": 3.98, + "learning_rate": 1.9999999999999998e-05, + "loss": 0.169, + "step": 84 + }, + { + "epoch": 3.98, + "eval_F1": 0.9385964912280702, + "eval_Precision": 0.9727272727272728, + "eval_Recall": 0.9067796610169492, + "eval_accuracy": 0.9591836734693877, + "eval_loss": 0.1545204222202301, + "eval_runtime": 1051.4657, + "eval_samples_per_second": 0.652, + "eval_steps_per_second": 0.021, + "step": 84 + }, + { + "epoch": 4.98, + "learning_rate": 1.6666666666666667e-05, + "loss": 0.1287, + "step": 105 + }, + { + "epoch": 4.98, + "eval_F1": 0.9406593406593408, + "eval_Precision": 0.9771689497716894, + "eval_Recall": 0.9067796610169492, + "eval_accuracy": 0.9606413994169096, + "eval_loss": 0.1248895600438118, + "eval_runtime": 1050.486, + "eval_samples_per_second": 0.653, + "eval_steps_per_second": 0.021, + "step": 105 + } + ], + "max_steps": 210, + "num_train_epochs": 10, + "total_flos": 6.845742942192e+17, + "trial_name": null, + "trial_params": null +} diff --git a/checkpoint-105/training_args.bin b/checkpoint-105/training_args.bin new file mode 100644 index 0000000000000000000000000000000000000000..44d4af4d6c45e0f8ad8da6879b0acf5f4add727f --- /dev/null +++ b/checkpoint-105/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4521181070c5206d39fa1359aeb62443676d9d133bf030a8d47fe7908ee32f60 +size 3503 diff --git a/checkpoint-126/config.json b/checkpoint-126/config.json new file mode 100644 index 0000000000000000000000000000000000000000..5895b668646a423372f040c92bb60d42a5449727 --- /dev/null +++ b/checkpoint-126/config.json @@ -0,0 +1,125 @@ +{ + "_name_or_path": "facebook/wav2vec2-base", + "activation_dropout": 0.0, + "adapter_kernel_size": 3, + "adapter_stride": 2, + "add_adapter": false, + "apply_spec_augment": true, + "architectures": [ + "Wav2Vec2ForSequenceClassification" + ], + "attention_dropout": 0.1, + "bos_token_id": 1, + "classifier_proj_size": 256, + "codevector_dim": 256, + "contrastive_logits_temperature": 0.1, + "conv_bias": false, + "conv_dim": [ + 512, + 512, + 512, + 512, + 512, + 512, + 512 + ], + "conv_kernel": [ + 10, + 3, + 3, + 3, + 3, + 2, + 2 + ], + "conv_stride": [ + 5, + 2, + 2, + 2, + 2, + 2, + 2 + ], + "ctc_loss_reduction": "sum", + "ctc_zero_infinity": false, + "diversity_loss_weight": 0.1, + "do_stable_layer_norm": false, + "eos_token_id": 2, + "feat_extract_activation": "gelu", + "feat_extract_norm": "group", + "feat_proj_dropout": 0.1, + "feat_quantizer_dropout": 0.0, + "final_dropout": 0.0, + "freeze_feat_extract_train": true, + "hidden_act": "gelu", + "hidden_dropout": 0.1, + "hidden_size": 768, + "id2label": { + "0": 0, + "1": 1 + }, + "initializer_range": 0.02, + "intermediate_size": 3072, + "label2id": { + "0": "0", + "1": "1" + }, + "layer_norm_eps": 1e-05, + "layerdrop": 0.0, + "mask_channel_length": 10, + "mask_channel_min_space": 1, + "mask_channel_other": 0.0, + "mask_channel_prob": 0.0, + "mask_channel_selection": "static", + "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_min_space": 1, + "mask_time_other": 0.0, + "mask_time_prob": 0.05, + "mask_time_selection": "static", + "model_type": "wav2vec2", + "no_mask_channel_overlap": false, + "no_mask_time_overlap": false, + "num_adapter_layers": 3, + "num_attention_heads": 12, + "num_codevector_groups": 2, + "num_codevectors_per_group": 320, + "num_conv_pos_embedding_groups": 16, + "num_conv_pos_embeddings": 128, + "num_feat_extract_layers": 7, + "num_hidden_layers": 12, + "num_negatives": 100, + "output_hidden_size": 768, + "pad_token_id": 0, + "proj_codevector_dim": 256, + "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": "4.26.0", + "use_weighted_layer_sum": false, + "vocab_size": 32, + "xvector_output_dim": 512 +} diff --git a/checkpoint-126/optimizer.pt b/checkpoint-126/optimizer.pt new file mode 100644 index 0000000000000000000000000000000000000000..4acbb7753964f72a2e982ba95b403e3fedcf6a8b --- /dev/null +++ b/checkpoint-126/optimizer.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:30404cc39fe15b70d457a896d04a54569e720e2d5c9c274382743bf24c9eb71d +size 756671269 diff --git a/checkpoint-126/preprocessor_config.json b/checkpoint-126/preprocessor_config.json new file mode 100644 index 0000000000000000000000000000000000000000..a0b7227fc1d916e469b14f6c154ad6dfea1e6891 --- /dev/null +++ b/checkpoint-126/preprocessor_config.json @@ -0,0 +1,9 @@ +{ + "do_normalize": true, + "feature_extractor_type": "Wav2Vec2FeatureExtractor", + "feature_size": 1, + "padding_side": "right", + "padding_value": 0.0, + "return_attention_mask": false, + "sampling_rate": 16000 +} diff --git a/checkpoint-126/pytorch_model.bin b/checkpoint-126/pytorch_model.bin new file mode 100644 index 0000000000000000000000000000000000000000..c38cc8ae2411054204f2323ee652270c5ac80ff3 --- /dev/null +++ b/checkpoint-126/pytorch_model.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f32c62d6949a5bc5028c602c846d2f6fb114156a3cd4a624c6bb5f6ec55cc6ba +size 378345555 diff --git a/checkpoint-126/rng_state.pth b/checkpoint-126/rng_state.pth new file mode 100644 index 0000000000000000000000000000000000000000..347f42df6c5c7c5522a59883331faf9fc933e68c --- /dev/null +++ b/checkpoint-126/rng_state.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3aeb08987691032a3643895d05d5edf2f416ffcb413a0d620f51c007acb89cf9 +size 13611 diff --git a/checkpoint-126/scheduler.pt b/checkpoint-126/scheduler.pt new file mode 100644 index 0000000000000000000000000000000000000000..e3d733d9608ac2121eb2a80b34123434398ea8a0 --- /dev/null +++ b/checkpoint-126/scheduler.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1c22e09d6e80e9e7c82d970fd2f046440b76865cb3bc483ea57adac8a2e0d013 +size 623 diff --git a/checkpoint-126/trainer_state.json b/checkpoint-126/trainer_state.json new file mode 100644 index 0000000000000000000000000000000000000000..0f21da08a2bb63daa6333cc6916377e01b2c940f --- /dev/null +++ b/checkpoint-126/trainer_state.json @@ -0,0 +1,130 @@ +{ + "best_metric": 0.9594882729211086, + "best_model_checkpoint": "wav2vec2-base-is_vinyl_scratched_or_not/checkpoint-126", + "epoch": 5.976744186046512, + "global_step": 126, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.05, + "learning_rate": 1.4285714285714286e-06, + "loss": 0.6901, + "step": 1 + }, + { + "epoch": 0.98, + "learning_rate": 3e-05, + "loss": 0.6671, + "step": 21 + }, + { + "epoch": 0.98, + "eval_F1": 0.0, + "eval_Precision": 0.0, + "eval_Recall": 0.0, + "eval_accuracy": 0.6559766763848397, + "eval_loss": 0.6235492825508118, + "eval_runtime": 1064.9816, + "eval_samples_per_second": 0.644, + "eval_steps_per_second": 0.021, + "step": 21 + }, + { + "epoch": 1.98, + "learning_rate": 2.6666666666666667e-05, + "loss": 0.4954, + "step": 42 + }, + { + "epoch": 1.98, + "eval_F1": 0.9095022624434389, + "eval_Precision": 0.9757281553398058, + "eval_Recall": 0.8516949152542372, + "eval_accuracy": 0.9416909620991254, + "eval_loss": 0.28244873881340027, + "eval_runtime": 1042.1602, + "eval_samples_per_second": 0.658, + "eval_steps_per_second": 0.021, + "step": 42 + }, + { + "epoch": 2.98, + "learning_rate": 2.3333333333333336e-05, + "loss": 0.2406, + "step": 63 + }, + { + "epoch": 2.98, + "eval_F1": 0.9336283185840708, + "eval_Precision": 0.9768518518518519, + "eval_Recall": 0.8940677966101694, + "eval_accuracy": 0.956268221574344, + "eval_loss": 0.17554502189159393, + "eval_runtime": 1040.1013, + "eval_samples_per_second": 0.66, + "eval_steps_per_second": 0.021, + "step": 63 + }, + { + "epoch": 3.98, + "learning_rate": 1.9999999999999998e-05, + "loss": 0.169, + "step": 84 + }, + { + "epoch": 3.98, + "eval_F1": 0.9385964912280702, + "eval_Precision": 0.9727272727272728, + "eval_Recall": 0.9067796610169492, + "eval_accuracy": 0.9591836734693877, + "eval_loss": 0.1545204222202301, + "eval_runtime": 1051.4657, + "eval_samples_per_second": 0.652, + "eval_steps_per_second": 0.021, + "step": 84 + }, + { + "epoch": 4.98, + "learning_rate": 1.6666666666666667e-05, + "loss": 0.1287, + "step": 105 + }, + { + "epoch": 4.98, + "eval_F1": 0.9406593406593408, + "eval_Precision": 0.9771689497716894, + "eval_Recall": 0.9067796610169492, + "eval_accuracy": 0.9606413994169096, + "eval_loss": 0.1248895600438118, + "eval_runtime": 1050.486, + "eval_samples_per_second": 0.653, + "eval_steps_per_second": 0.021, + "step": 105 + }, + { + "epoch": 5.98, + "learning_rate": 1.3333333333333333e-05, + "loss": 0.1102, + "step": 126 + }, + { + "epoch": 5.98, + "eval_F1": 0.9594882729211086, + "eval_Precision": 0.9656652360515021, + "eval_Recall": 0.9533898305084746, + "eval_accuracy": 0.9723032069970845, + "eval_loss": 0.11587227135896683, + "eval_runtime": 1049.2012, + "eval_samples_per_second": 0.654, + "eval_steps_per_second": 0.021, + "step": 126 + } + ], + "max_steps": 210, + "num_train_epochs": 10, + "total_flos": 8.2148915306304e+17, + "trial_name": null, + "trial_params": null +} diff --git a/checkpoint-126/training_args.bin b/checkpoint-126/training_args.bin new file mode 100644 index 0000000000000000000000000000000000000000..44d4af4d6c45e0f8ad8da6879b0acf5f4add727f --- /dev/null +++ b/checkpoint-126/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4521181070c5206d39fa1359aeb62443676d9d133bf030a8d47fe7908ee32f60 +size 3503 diff --git a/checkpoint-147/config.json b/checkpoint-147/config.json new file mode 100644 index 0000000000000000000000000000000000000000..5895b668646a423372f040c92bb60d42a5449727 --- /dev/null +++ b/checkpoint-147/config.json @@ -0,0 +1,125 @@ +{ + "_name_or_path": "facebook/wav2vec2-base", + "activation_dropout": 0.0, + "adapter_kernel_size": 3, + "adapter_stride": 2, + "add_adapter": false, + "apply_spec_augment": true, + "architectures": [ + "Wav2Vec2ForSequenceClassification" + ], + "attention_dropout": 0.1, + "bos_token_id": 1, + "classifier_proj_size": 256, + "codevector_dim": 256, + "contrastive_logits_temperature": 0.1, + "conv_bias": false, + "conv_dim": [ + 512, + 512, + 512, + 512, + 512, + 512, + 512 + ], + "conv_kernel": [ + 10, + 3, + 3, + 3, + 3, + 2, + 2 + ], + "conv_stride": [ + 5, + 2, + 2, + 2, + 2, + 2, + 2 + ], + "ctc_loss_reduction": "sum", + "ctc_zero_infinity": false, + "diversity_loss_weight": 0.1, + "do_stable_layer_norm": false, + "eos_token_id": 2, + "feat_extract_activation": "gelu", + "feat_extract_norm": "group", + "feat_proj_dropout": 0.1, + "feat_quantizer_dropout": 0.0, + "final_dropout": 0.0, + "freeze_feat_extract_train": true, + "hidden_act": "gelu", + "hidden_dropout": 0.1, + "hidden_size": 768, + "id2label": { + "0": 0, + "1": 1 + }, + "initializer_range": 0.02, + "intermediate_size": 3072, + "label2id": { + "0": "0", + "1": "1" + }, + "layer_norm_eps": 1e-05, + "layerdrop": 0.0, + "mask_channel_length": 10, + "mask_channel_min_space": 1, + "mask_channel_other": 0.0, + "mask_channel_prob": 0.0, + "mask_channel_selection": "static", + "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_min_space": 1, + "mask_time_other": 0.0, + "mask_time_prob": 0.05, + "mask_time_selection": "static", + "model_type": "wav2vec2", + "no_mask_channel_overlap": false, + "no_mask_time_overlap": false, + "num_adapter_layers": 3, + "num_attention_heads": 12, + "num_codevector_groups": 2, + "num_codevectors_per_group": 320, + "num_conv_pos_embedding_groups": 16, + "num_conv_pos_embeddings": 128, + "num_feat_extract_layers": 7, + "num_hidden_layers": 12, + "num_negatives": 100, + "output_hidden_size": 768, + "pad_token_id": 0, + "proj_codevector_dim": 256, + "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": "4.26.0", + "use_weighted_layer_sum": false, + "vocab_size": 32, + "xvector_output_dim": 512 +} diff --git a/checkpoint-147/optimizer.pt b/checkpoint-147/optimizer.pt new file mode 100644 index 0000000000000000000000000000000000000000..bb68dd5f6e438ccc3fb6a956bff842b482487ff0 --- /dev/null +++ b/checkpoint-147/optimizer.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:cd8792a4eb875c1716cddbb0b8ef83d84c280350a6dd26c8e072b8fd251d1f04 +size 756671269 diff --git a/checkpoint-147/preprocessor_config.json b/checkpoint-147/preprocessor_config.json new file mode 100644 index 0000000000000000000000000000000000000000..a0b7227fc1d916e469b14f6c154ad6dfea1e6891 --- /dev/null +++ b/checkpoint-147/preprocessor_config.json @@ -0,0 +1,9 @@ +{ + "do_normalize": true, + "feature_extractor_type": "Wav2Vec2FeatureExtractor", + "feature_size": 1, + "padding_side": "right", + "padding_value": 0.0, + "return_attention_mask": false, + "sampling_rate": 16000 +} diff --git a/checkpoint-147/pytorch_model.bin b/checkpoint-147/pytorch_model.bin new file mode 100644 index 0000000000000000000000000000000000000000..9a2e6432086753a23e91858749cfeaa48948fdd0 --- /dev/null +++ b/checkpoint-147/pytorch_model.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ff3255ba90fd5778c017bd790e559b1444d916354bf63a360fc7db43c435cd83 +size 378345555 diff --git a/checkpoint-147/rng_state.pth b/checkpoint-147/rng_state.pth new file mode 100644 index 0000000000000000000000000000000000000000..a8a4338bc6d07feac70fd94112476921cfea72ba --- /dev/null +++ b/checkpoint-147/rng_state.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b82889f72e9194f8f2e54e48553d15a3facdacb0c7e99f06a72b71f2e4f3fb24 +size 13611 diff --git a/checkpoint-147/scheduler.pt b/checkpoint-147/scheduler.pt new file mode 100644 index 0000000000000000000000000000000000000000..7e293d3f2629b0ade87233a4e53a3cfd18e12bc4 --- /dev/null +++ b/checkpoint-147/scheduler.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:40f575b7a521d42834c25781b03ebcb41b842dae0e69331794ec5b689a01173a +size 623 diff --git a/checkpoint-147/trainer_state.json b/checkpoint-147/trainer_state.json new file mode 100644 index 0000000000000000000000000000000000000000..3530a193164bf2301f803a59a293936212e032aa --- /dev/null +++ b/checkpoint-147/trainer_state.json @@ -0,0 +1,148 @@ +{ + "best_metric": 0.9594882729211086, + "best_model_checkpoint": "wav2vec2-base-is_vinyl_scratched_or_not/checkpoint-126", + "epoch": 6.976744186046512, + "global_step": 147, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.05, + "learning_rate": 1.4285714285714286e-06, + "loss": 0.6901, + "step": 1 + }, + { + "epoch": 0.98, + "learning_rate": 3e-05, + "loss": 0.6671, + "step": 21 + }, + { + "epoch": 0.98, + "eval_F1": 0.0, + "eval_Precision": 0.0, + "eval_Recall": 0.0, + "eval_accuracy": 0.6559766763848397, + "eval_loss": 0.6235492825508118, + "eval_runtime": 1064.9816, + "eval_samples_per_second": 0.644, + "eval_steps_per_second": 0.021, + "step": 21 + }, + { + "epoch": 1.98, + "learning_rate": 2.6666666666666667e-05, + "loss": 0.4954, + "step": 42 + }, + { + "epoch": 1.98, + "eval_F1": 0.9095022624434389, + "eval_Precision": 0.9757281553398058, + "eval_Recall": 0.8516949152542372, + "eval_accuracy": 0.9416909620991254, + "eval_loss": 0.28244873881340027, + "eval_runtime": 1042.1602, + "eval_samples_per_second": 0.658, + "eval_steps_per_second": 0.021, + "step": 42 + }, + { + "epoch": 2.98, + "learning_rate": 2.3333333333333336e-05, + "loss": 0.2406, + "step": 63 + }, + { + "epoch": 2.98, + "eval_F1": 0.9336283185840708, + "eval_Precision": 0.9768518518518519, + "eval_Recall": 0.8940677966101694, + "eval_accuracy": 0.956268221574344, + "eval_loss": 0.17554502189159393, + "eval_runtime": 1040.1013, + "eval_samples_per_second": 0.66, + "eval_steps_per_second": 0.021, + "step": 63 + }, + { + "epoch": 3.98, + "learning_rate": 1.9999999999999998e-05, + "loss": 0.169, + "step": 84 + }, + { + "epoch": 3.98, + "eval_F1": 0.9385964912280702, + "eval_Precision": 0.9727272727272728, + "eval_Recall": 0.9067796610169492, + "eval_accuracy": 0.9591836734693877, + "eval_loss": 0.1545204222202301, + "eval_runtime": 1051.4657, + "eval_samples_per_second": 0.652, + "eval_steps_per_second": 0.021, + "step": 84 + }, + { + "epoch": 4.98, + "learning_rate": 1.6666666666666667e-05, + "loss": 0.1287, + "step": 105 + }, + { + "epoch": 4.98, + "eval_F1": 0.9406593406593408, + "eval_Precision": 0.9771689497716894, + "eval_Recall": 0.9067796610169492, + "eval_accuracy": 0.9606413994169096, + "eval_loss": 0.1248895600438118, + "eval_runtime": 1050.486, + "eval_samples_per_second": 0.653, + "eval_steps_per_second": 0.021, + "step": 105 + }, + { + "epoch": 5.98, + "learning_rate": 1.3333333333333333e-05, + "loss": 0.1102, + "step": 126 + }, + { + "epoch": 5.98, + "eval_F1": 0.9594882729211086, + "eval_Precision": 0.9656652360515021, + "eval_Recall": 0.9533898305084746, + "eval_accuracy": 0.9723032069970845, + "eval_loss": 0.11587227135896683, + "eval_runtime": 1049.2012, + "eval_samples_per_second": 0.654, + "eval_steps_per_second": 0.021, + "step": 126 + }, + { + "epoch": 6.98, + "learning_rate": 9.999999999999999e-06, + "loss": 0.0923, + "step": 147 + }, + { + "epoch": 6.98, + "eval_F1": 0.9515789473684211, + "eval_Precision": 0.9456066945606695, + "eval_Recall": 0.9576271186440678, + "eval_accuracy": 0.9664723032069971, + "eval_loss": 0.10729651153087616, + "eval_runtime": 1059.4975, + "eval_samples_per_second": 0.647, + "eval_steps_per_second": 0.021, + "step": 147 + } + ], + "max_steps": 210, + "num_train_epochs": 10, + "total_flos": 9.5840401190688e+17, + "trial_name": null, + "trial_params": null +} diff --git a/checkpoint-147/training_args.bin b/checkpoint-147/training_args.bin new file mode 100644 index 0000000000000000000000000000000000000000..44d4af4d6c45e0f8ad8da6879b0acf5f4add727f --- /dev/null +++ b/checkpoint-147/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4521181070c5206d39fa1359aeb62443676d9d133bf030a8d47fe7908ee32f60 +size 3503 diff --git a/checkpoint-168/config.json b/checkpoint-168/config.json new file mode 100644 index 0000000000000000000000000000000000000000..5895b668646a423372f040c92bb60d42a5449727 --- /dev/null +++ b/checkpoint-168/config.json @@ -0,0 +1,125 @@ +{ + "_name_or_path": "facebook/wav2vec2-base", + "activation_dropout": 0.0, + "adapter_kernel_size": 3, + "adapter_stride": 2, + "add_adapter": false, + "apply_spec_augment": true, + "architectures": [ + "Wav2Vec2ForSequenceClassification" + ], + "attention_dropout": 0.1, + "bos_token_id": 1, + "classifier_proj_size": 256, + "codevector_dim": 256, + "contrastive_logits_temperature": 0.1, + "conv_bias": false, + "conv_dim": [ + 512, + 512, + 512, + 512, + 512, + 512, + 512 + ], + "conv_kernel": [ + 10, + 3, + 3, + 3, + 3, + 2, + 2 + ], + "conv_stride": [ + 5, + 2, + 2, + 2, + 2, + 2, + 2 + ], + "ctc_loss_reduction": "sum", + "ctc_zero_infinity": false, + "diversity_loss_weight": 0.1, + "do_stable_layer_norm": false, + "eos_token_id": 2, + "feat_extract_activation": "gelu", + "feat_extract_norm": "group", + "feat_proj_dropout": 0.1, + "feat_quantizer_dropout": 0.0, + "final_dropout": 0.0, + "freeze_feat_extract_train": true, + "hidden_act": "gelu", + "hidden_dropout": 0.1, + "hidden_size": 768, + "id2label": { + "0": 0, + "1": 1 + }, + "initializer_range": 0.02, + "intermediate_size": 3072, + "label2id": { + "0": "0", + "1": "1" + }, + "layer_norm_eps": 1e-05, + "layerdrop": 0.0, + "mask_channel_length": 10, + "mask_channel_min_space": 1, + "mask_channel_other": 0.0, + "mask_channel_prob": 0.0, + "mask_channel_selection": "static", + "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_min_space": 1, + "mask_time_other": 0.0, + "mask_time_prob": 0.05, + "mask_time_selection": "static", + "model_type": "wav2vec2", + "no_mask_channel_overlap": false, + "no_mask_time_overlap": false, + "num_adapter_layers": 3, + "num_attention_heads": 12, + "num_codevector_groups": 2, + "num_codevectors_per_group": 320, + "num_conv_pos_embedding_groups": 16, + "num_conv_pos_embeddings": 128, + "num_feat_extract_layers": 7, + "num_hidden_layers": 12, + "num_negatives": 100, + "output_hidden_size": 768, + "pad_token_id": 0, + "proj_codevector_dim": 256, + "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": "4.26.0", + "use_weighted_layer_sum": false, + "vocab_size": 32, + "xvector_output_dim": 512 +} diff --git a/checkpoint-168/optimizer.pt b/checkpoint-168/optimizer.pt new file mode 100644 index 0000000000000000000000000000000000000000..9cb1599ffd116e9faefea8acaaa55dfae662dd8f --- /dev/null +++ b/checkpoint-168/optimizer.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2c3efb950e1d52afb01b74807bd334aadda26ae9db22d6814e2dd8340f4415f3 +size 756671269 diff --git a/checkpoint-168/preprocessor_config.json b/checkpoint-168/preprocessor_config.json new file mode 100644 index 0000000000000000000000000000000000000000..a0b7227fc1d916e469b14f6c154ad6dfea1e6891 --- /dev/null +++ b/checkpoint-168/preprocessor_config.json @@ -0,0 +1,9 @@ +{ + "do_normalize": true, + "feature_extractor_type": "Wav2Vec2FeatureExtractor", + "feature_size": 1, + "padding_side": "right", + "padding_value": 0.0, + "return_attention_mask": false, + "sampling_rate": 16000 +} diff --git a/checkpoint-168/pytorch_model.bin b/checkpoint-168/pytorch_model.bin new file mode 100644 index 0000000000000000000000000000000000000000..529b102bf9e63664b07f7465cca4bf33eee5c841 --- /dev/null +++ b/checkpoint-168/pytorch_model.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2d80feebf75b9add887b3409fea764a26b226005fc477c13b47098cca4579b04 +size 378345555 diff --git a/checkpoint-168/rng_state.pth b/checkpoint-168/rng_state.pth new file mode 100644 index 0000000000000000000000000000000000000000..e8873fb347ae6dcbc3bbe7e2ee5530e3d813a900 --- /dev/null +++ b/checkpoint-168/rng_state.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f4348501b22a26ac1aeb2e86bd788289116344a48402648de3777819ec84c429 +size 13611 diff --git a/checkpoint-168/scheduler.pt b/checkpoint-168/scheduler.pt new file mode 100644 index 0000000000000000000000000000000000000000..7224e91db50b7c6b48942e0e918bed64999d093a --- /dev/null +++ b/checkpoint-168/scheduler.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4660b0d5db40788f4a079cb632965b97fb6cb598256ca600a122531f32b5c711 +size 623 diff --git a/checkpoint-168/trainer_state.json b/checkpoint-168/trainer_state.json new file mode 100644 index 0000000000000000000000000000000000000000..d710f56ea3608ac6e65b78c27b677e25dc0be1da --- /dev/null +++ b/checkpoint-168/trainer_state.json @@ -0,0 +1,166 @@ +{ + "best_metric": 0.9637526652452025, + "best_model_checkpoint": "wav2vec2-base-is_vinyl_scratched_or_not/checkpoint-168", + "epoch": 7.976744186046512, + "global_step": 168, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.05, + "learning_rate": 1.4285714285714286e-06, + "loss": 0.6901, + "step": 1 + }, + { + "epoch": 0.98, + "learning_rate": 3e-05, + "loss": 0.6671, + "step": 21 + }, + { + "epoch": 0.98, + "eval_F1": 0.0, + "eval_Precision": 0.0, + "eval_Recall": 0.0, + "eval_accuracy": 0.6559766763848397, + "eval_loss": 0.6235492825508118, + "eval_runtime": 1064.9816, + "eval_samples_per_second": 0.644, + "eval_steps_per_second": 0.021, + "step": 21 + }, + { + "epoch": 1.98, + "learning_rate": 2.6666666666666667e-05, + "loss": 0.4954, + "step": 42 + }, + { + "epoch": 1.98, + "eval_F1": 0.9095022624434389, + "eval_Precision": 0.9757281553398058, + "eval_Recall": 0.8516949152542372, + "eval_accuracy": 0.9416909620991254, + "eval_loss": 0.28244873881340027, + "eval_runtime": 1042.1602, + "eval_samples_per_second": 0.658, + "eval_steps_per_second": 0.021, + "step": 42 + }, + { + "epoch": 2.98, + "learning_rate": 2.3333333333333336e-05, + "loss": 0.2406, + "step": 63 + }, + { + "epoch": 2.98, + "eval_F1": 0.9336283185840708, + "eval_Precision": 0.9768518518518519, + "eval_Recall": 0.8940677966101694, + "eval_accuracy": 0.956268221574344, + "eval_loss": 0.17554502189159393, + "eval_runtime": 1040.1013, + "eval_samples_per_second": 0.66, + "eval_steps_per_second": 0.021, + "step": 63 + }, + { + "epoch": 3.98, + "learning_rate": 1.9999999999999998e-05, + "loss": 0.169, + "step": 84 + }, + { + "epoch": 3.98, + "eval_F1": 0.9385964912280702, + "eval_Precision": 0.9727272727272728, + "eval_Recall": 0.9067796610169492, + "eval_accuracy": 0.9591836734693877, + "eval_loss": 0.1545204222202301, + "eval_runtime": 1051.4657, + "eval_samples_per_second": 0.652, + "eval_steps_per_second": 0.021, + "step": 84 + }, + { + "epoch": 4.98, + "learning_rate": 1.6666666666666667e-05, + "loss": 0.1287, + "step": 105 + }, + { + "epoch": 4.98, + "eval_F1": 0.9406593406593408, + "eval_Precision": 0.9771689497716894, + "eval_Recall": 0.9067796610169492, + "eval_accuracy": 0.9606413994169096, + "eval_loss": 0.1248895600438118, + "eval_runtime": 1050.486, + "eval_samples_per_second": 0.653, + "eval_steps_per_second": 0.021, + "step": 105 + }, + { + "epoch": 5.98, + "learning_rate": 1.3333333333333333e-05, + "loss": 0.1102, + "step": 126 + }, + { + "epoch": 5.98, + "eval_F1": 0.9594882729211086, + "eval_Precision": 0.9656652360515021, + "eval_Recall": 0.9533898305084746, + "eval_accuracy": 0.9723032069970845, + "eval_loss": 0.11587227135896683, + "eval_runtime": 1049.2012, + "eval_samples_per_second": 0.654, + "eval_steps_per_second": 0.021, + "step": 126 + }, + { + "epoch": 6.98, + "learning_rate": 9.999999999999999e-06, + "loss": 0.0923, + "step": 147 + }, + { + "epoch": 6.98, + "eval_F1": 0.9515789473684211, + "eval_Precision": 0.9456066945606695, + "eval_Recall": 0.9576271186440678, + "eval_accuracy": 0.9664723032069971, + "eval_loss": 0.10729651153087616, + "eval_runtime": 1059.4975, + "eval_samples_per_second": 0.647, + "eval_steps_per_second": 0.021, + "step": 147 + }, + { + "epoch": 7.98, + "learning_rate": 6.666666666666667e-06, + "loss": 0.0877, + "step": 168 + }, + { + "epoch": 7.98, + "eval_F1": 0.9637526652452025, + "eval_Precision": 0.9699570815450643, + "eval_Recall": 0.9576271186440678, + "eval_accuracy": 0.9752186588921283, + "eval_loss": 0.10392837971448898, + "eval_runtime": 1046.4548, + "eval_samples_per_second": 0.656, + "eval_steps_per_second": 0.021, + "step": 168 + } + ], + "max_steps": 210, + "num_train_epochs": 10, + "total_flos": 1.09531887075072e+18, + "trial_name": null, + "trial_params": null +} diff --git a/checkpoint-168/training_args.bin b/checkpoint-168/training_args.bin new file mode 100644 index 0000000000000000000000000000000000000000..44d4af4d6c45e0f8ad8da6879b0acf5f4add727f --- /dev/null +++ b/checkpoint-168/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4521181070c5206d39fa1359aeb62443676d9d133bf030a8d47fe7908ee32f60 +size 3503 diff --git a/checkpoint-189/config.json b/checkpoint-189/config.json new file mode 100644 index 0000000000000000000000000000000000000000..5895b668646a423372f040c92bb60d42a5449727 --- /dev/null +++ b/checkpoint-189/config.json @@ -0,0 +1,125 @@ +{ + "_name_or_path": "facebook/wav2vec2-base", + "activation_dropout": 0.0, + "adapter_kernel_size": 3, + "adapter_stride": 2, + "add_adapter": false, + "apply_spec_augment": true, + "architectures": [ + "Wav2Vec2ForSequenceClassification" + ], + "attention_dropout": 0.1, + "bos_token_id": 1, + "classifier_proj_size": 256, + "codevector_dim": 256, + "contrastive_logits_temperature": 0.1, + "conv_bias": false, + "conv_dim": [ + 512, + 512, + 512, + 512, + 512, + 512, + 512 + ], + "conv_kernel": [ + 10, + 3, + 3, + 3, + 3, + 2, + 2 + ], + "conv_stride": [ + 5, + 2, + 2, + 2, + 2, + 2, + 2 + ], + "ctc_loss_reduction": "sum", + "ctc_zero_infinity": false, + "diversity_loss_weight": 0.1, + "do_stable_layer_norm": false, + "eos_token_id": 2, + "feat_extract_activation": "gelu", + "feat_extract_norm": "group", + "feat_proj_dropout": 0.1, + "feat_quantizer_dropout": 0.0, + "final_dropout": 0.0, + "freeze_feat_extract_train": true, + "hidden_act": "gelu", + "hidden_dropout": 0.1, + "hidden_size": 768, + "id2label": { + "0": 0, + "1": 1 + }, + "initializer_range": 0.02, + "intermediate_size": 3072, + "label2id": { + "0": "0", + "1": "1" + }, + "layer_norm_eps": 1e-05, + "layerdrop": 0.0, + "mask_channel_length": 10, + "mask_channel_min_space": 1, + "mask_channel_other": 0.0, + "mask_channel_prob": 0.0, + "mask_channel_selection": "static", + "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_min_space": 1, + "mask_time_other": 0.0, + "mask_time_prob": 0.05, + "mask_time_selection": "static", + "model_type": "wav2vec2", + "no_mask_channel_overlap": false, + "no_mask_time_overlap": false, + "num_adapter_layers": 3, + "num_attention_heads": 12, + "num_codevector_groups": 2, + "num_codevectors_per_group": 320, + "num_conv_pos_embedding_groups": 16, + "num_conv_pos_embeddings": 128, + "num_feat_extract_layers": 7, + "num_hidden_layers": 12, + "num_negatives": 100, + "output_hidden_size": 768, + "pad_token_id": 0, + "proj_codevector_dim": 256, + "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": "4.26.0", + "use_weighted_layer_sum": false, + "vocab_size": 32, + "xvector_output_dim": 512 +} diff --git a/checkpoint-189/optimizer.pt b/checkpoint-189/optimizer.pt new file mode 100644 index 0000000000000000000000000000000000000000..d4e6e5a2d9802aef288f2e64ee875c7849be6dbc --- /dev/null +++ b/checkpoint-189/optimizer.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:79ef3cb8889aaaab5a453bf47370bf98789ed72760255063c98e437298875d82 +size 756671269 diff --git a/checkpoint-189/preprocessor_config.json b/checkpoint-189/preprocessor_config.json new file mode 100644 index 0000000000000000000000000000000000000000..a0b7227fc1d916e469b14f6c154ad6dfea1e6891 --- /dev/null +++ b/checkpoint-189/preprocessor_config.json @@ -0,0 +1,9 @@ +{ + "do_normalize": true, + "feature_extractor_type": "Wav2Vec2FeatureExtractor", + "feature_size": 1, + "padding_side": "right", + "padding_value": 0.0, + "return_attention_mask": false, + "sampling_rate": 16000 +} diff --git a/checkpoint-189/pytorch_model.bin b/checkpoint-189/pytorch_model.bin new file mode 100644 index 0000000000000000000000000000000000000000..4ceffc949a037ef7635063184932555d5f0eb306 --- /dev/null +++ b/checkpoint-189/pytorch_model.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8657ad0496958f12263ab116c4668fecd090d248dad239bb4725f83bb444bcb2 +size 378345555 diff --git a/checkpoint-189/rng_state.pth b/checkpoint-189/rng_state.pth new file mode 100644 index 0000000000000000000000000000000000000000..fc73dd5adb62439bd78d1c57c06557f0598321b4 --- /dev/null +++ b/checkpoint-189/rng_state.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:72e80e4564843ebfeac09ae18cd3b55e53b0138af482143fcd87c1db6d4d2623 +size 13611 diff --git a/checkpoint-189/scheduler.pt b/checkpoint-189/scheduler.pt new file mode 100644 index 0000000000000000000000000000000000000000..ae473a6e04824a6fb3a7b4ab5b29f60f4ed9bc01 --- /dev/null +++ b/checkpoint-189/scheduler.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3e491ae4f90e3635a94ed7f571a9260440304eccbe49682f12a95646aaeb21c4 +size 623 diff --git a/checkpoint-189/trainer_state.json b/checkpoint-189/trainer_state.json new file mode 100644 index 0000000000000000000000000000000000000000..237f0fccca286170e4f32623d72523d5e3822b57 --- /dev/null +++ b/checkpoint-189/trainer_state.json @@ -0,0 +1,184 @@ +{ + "best_metric": 0.9637526652452025, + "best_model_checkpoint": "wav2vec2-base-is_vinyl_scratched_or_not/checkpoint-168", + "epoch": 8.976744186046512, + "global_step": 189, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.05, + "learning_rate": 1.4285714285714286e-06, + "loss": 0.6901, + "step": 1 + }, + { + "epoch": 0.98, + "learning_rate": 3e-05, + "loss": 0.6671, + "step": 21 + }, + { + "epoch": 0.98, + "eval_F1": 0.0, + "eval_Precision": 0.0, + "eval_Recall": 0.0, + "eval_accuracy": 0.6559766763848397, + "eval_loss": 0.6235492825508118, + "eval_runtime": 1064.9816, + "eval_samples_per_second": 0.644, + "eval_steps_per_second": 0.021, + "step": 21 + }, + { + "epoch": 1.98, + "learning_rate": 2.6666666666666667e-05, + "loss": 0.4954, + "step": 42 + }, + { + "epoch": 1.98, + "eval_F1": 0.9095022624434389, + "eval_Precision": 0.9757281553398058, + "eval_Recall": 0.8516949152542372, + "eval_accuracy": 0.9416909620991254, + "eval_loss": 0.28244873881340027, + "eval_runtime": 1042.1602, + "eval_samples_per_second": 0.658, + "eval_steps_per_second": 0.021, + "step": 42 + }, + { + "epoch": 2.98, + "learning_rate": 2.3333333333333336e-05, + "loss": 0.2406, + "step": 63 + }, + { + "epoch": 2.98, + "eval_F1": 0.9336283185840708, + "eval_Precision": 0.9768518518518519, + "eval_Recall": 0.8940677966101694, + "eval_accuracy": 0.956268221574344, + "eval_loss": 0.17554502189159393, + "eval_runtime": 1040.1013, + "eval_samples_per_second": 0.66, + "eval_steps_per_second": 0.021, + "step": 63 + }, + { + "epoch": 3.98, + "learning_rate": 1.9999999999999998e-05, + "loss": 0.169, + "step": 84 + }, + { + "epoch": 3.98, + "eval_F1": 0.9385964912280702, + "eval_Precision": 0.9727272727272728, + "eval_Recall": 0.9067796610169492, + "eval_accuracy": 0.9591836734693877, + "eval_loss": 0.1545204222202301, + "eval_runtime": 1051.4657, + "eval_samples_per_second": 0.652, + "eval_steps_per_second": 0.021, + "step": 84 + }, + { + "epoch": 4.98, + "learning_rate": 1.6666666666666667e-05, + "loss": 0.1287, + "step": 105 + }, + { + "epoch": 4.98, + "eval_F1": 0.9406593406593408, + "eval_Precision": 0.9771689497716894, + "eval_Recall": 0.9067796610169492, + "eval_accuracy": 0.9606413994169096, + "eval_loss": 0.1248895600438118, + "eval_runtime": 1050.486, + "eval_samples_per_second": 0.653, + "eval_steps_per_second": 0.021, + "step": 105 + }, + { + "epoch": 5.98, + "learning_rate": 1.3333333333333333e-05, + "loss": 0.1102, + "step": 126 + }, + { + "epoch": 5.98, + "eval_F1": 0.9594882729211086, + "eval_Precision": 0.9656652360515021, + "eval_Recall": 0.9533898305084746, + "eval_accuracy": 0.9723032069970845, + "eval_loss": 0.11587227135896683, + "eval_runtime": 1049.2012, + "eval_samples_per_second": 0.654, + "eval_steps_per_second": 0.021, + "step": 126 + }, + { + "epoch": 6.98, + "learning_rate": 9.999999999999999e-06, + "loss": 0.0923, + "step": 147 + }, + { + "epoch": 6.98, + "eval_F1": 0.9515789473684211, + "eval_Precision": 0.9456066945606695, + "eval_Recall": 0.9576271186440678, + "eval_accuracy": 0.9664723032069971, + "eval_loss": 0.10729651153087616, + "eval_runtime": 1059.4975, + "eval_samples_per_second": 0.647, + "eval_steps_per_second": 0.021, + "step": 147 + }, + { + "epoch": 7.98, + "learning_rate": 6.666666666666667e-06, + "loss": 0.0877, + "step": 168 + }, + { + "epoch": 7.98, + "eval_F1": 0.9637526652452025, + "eval_Precision": 0.9699570815450643, + "eval_Recall": 0.9576271186440678, + "eval_accuracy": 0.9752186588921283, + "eval_loss": 0.10392837971448898, + "eval_runtime": 1046.4548, + "eval_samples_per_second": 0.656, + "eval_steps_per_second": 0.021, + "step": 168 + }, + { + "epoch": 8.98, + "learning_rate": 3.3333333333333333e-06, + "loss": 0.0807, + "step": 189 + }, + { + "epoch": 8.98, + "eval_F1": 0.9535864978902954, + "eval_Precision": 0.9495798319327731, + "eval_Recall": 0.9576271186440678, + "eval_accuracy": 0.967930029154519, + "eval_loss": 0.1087886393070221, + "eval_runtime": 1049.1807, + "eval_samples_per_second": 0.654, + "eval_steps_per_second": 0.021, + "step": 189 + } + ], + "max_steps": 210, + "num_train_epochs": 10, + "total_flos": 1.23223372959456e+18, + "trial_name": null, + "trial_params": null +} diff --git a/checkpoint-189/training_args.bin b/checkpoint-189/training_args.bin new file mode 100644 index 0000000000000000000000000000000000000000..44d4af4d6c45e0f8ad8da6879b0acf5f4add727f --- /dev/null +++ b/checkpoint-189/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4521181070c5206d39fa1359aeb62443676d9d133bf030a8d47fe7908ee32f60 +size 3503 diff --git a/checkpoint-21/config.json b/checkpoint-21/config.json new file mode 100644 index 0000000000000000000000000000000000000000..5895b668646a423372f040c92bb60d42a5449727 --- /dev/null +++ b/checkpoint-21/config.json @@ -0,0 +1,125 @@ +{ + "_name_or_path": "facebook/wav2vec2-base", + "activation_dropout": 0.0, + "adapter_kernel_size": 3, + "adapter_stride": 2, + "add_adapter": false, + "apply_spec_augment": true, + "architectures": [ + "Wav2Vec2ForSequenceClassification" + ], + "attention_dropout": 0.1, + "bos_token_id": 1, + "classifier_proj_size": 256, + "codevector_dim": 256, + "contrastive_logits_temperature": 0.1, + "conv_bias": false, + "conv_dim": [ + 512, + 512, + 512, + 512, + 512, + 512, + 512 + ], + "conv_kernel": [ + 10, + 3, + 3, + 3, + 3, + 2, + 2 + ], + "conv_stride": [ + 5, + 2, + 2, + 2, + 2, + 2, + 2 + ], + "ctc_loss_reduction": "sum", + "ctc_zero_infinity": false, + "diversity_loss_weight": 0.1, + "do_stable_layer_norm": false, + "eos_token_id": 2, + "feat_extract_activation": "gelu", + "feat_extract_norm": "group", + "feat_proj_dropout": 0.1, + "feat_quantizer_dropout": 0.0, + "final_dropout": 0.0, + "freeze_feat_extract_train": true, + "hidden_act": "gelu", + "hidden_dropout": 0.1, + "hidden_size": 768, + "id2label": { + "0": 0, + "1": 1 + }, + "initializer_range": 0.02, + "intermediate_size": 3072, + "label2id": { + "0": "0", + "1": "1" + }, + "layer_norm_eps": 1e-05, + "layerdrop": 0.0, + "mask_channel_length": 10, + "mask_channel_min_space": 1, + "mask_channel_other": 0.0, + "mask_channel_prob": 0.0, + "mask_channel_selection": "static", + "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_min_space": 1, + "mask_time_other": 0.0, + "mask_time_prob": 0.05, + "mask_time_selection": "static", + "model_type": "wav2vec2", + "no_mask_channel_overlap": false, + "no_mask_time_overlap": false, + "num_adapter_layers": 3, + "num_attention_heads": 12, + "num_codevector_groups": 2, + "num_codevectors_per_group": 320, + "num_conv_pos_embedding_groups": 16, + "num_conv_pos_embeddings": 128, + "num_feat_extract_layers": 7, + "num_hidden_layers": 12, + "num_negatives": 100, + "output_hidden_size": 768, + "pad_token_id": 0, + "proj_codevector_dim": 256, + "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": "4.26.0", + "use_weighted_layer_sum": false, + "vocab_size": 32, + "xvector_output_dim": 512 +} diff --git a/checkpoint-21/optimizer.pt b/checkpoint-21/optimizer.pt new file mode 100644 index 0000000000000000000000000000000000000000..034bd8c5926baa2e30e8caa319179b0b9faab6a9 --- /dev/null +++ b/checkpoint-21/optimizer.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a4d717d50ea851bbd0760c0ae423052eefdaee77dd648084a79f044065794371 +size 756671269 diff --git a/checkpoint-21/preprocessor_config.json b/checkpoint-21/preprocessor_config.json new file mode 100644 index 0000000000000000000000000000000000000000..a0b7227fc1d916e469b14f6c154ad6dfea1e6891 --- /dev/null +++ b/checkpoint-21/preprocessor_config.json @@ -0,0 +1,9 @@ +{ + "do_normalize": true, + "feature_extractor_type": "Wav2Vec2FeatureExtractor", + "feature_size": 1, + "padding_side": "right", + "padding_value": 0.0, + "return_attention_mask": false, + "sampling_rate": 16000 +} diff --git a/checkpoint-21/pytorch_model.bin b/checkpoint-21/pytorch_model.bin new file mode 100644 index 0000000000000000000000000000000000000000..b4633fa678c47e0b8e73752ceecde728ecbf7f0c --- /dev/null +++ b/checkpoint-21/pytorch_model.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c0371786eca2dbe97f5e2252e9fde629e17ec822d1238935c2d5925d3698ea23 +size 378345555 diff --git a/checkpoint-21/rng_state.pth b/checkpoint-21/rng_state.pth new file mode 100644 index 0000000000000000000000000000000000000000..1a2b9f0f85da9dcc9c6eaf7cb7ffabbbfff5e106 --- /dev/null +++ b/checkpoint-21/rng_state.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f2e53348e6c5c6bfde5399afe15ee8fc9cc0fe4f71f4fcaf81a61b7ef10eccb9 +size 13611 diff --git a/checkpoint-21/scheduler.pt b/checkpoint-21/scheduler.pt new file mode 100644 index 0000000000000000000000000000000000000000..917fc6b48fc33645e8dcc28b8407f6d907f9e266 --- /dev/null +++ b/checkpoint-21/scheduler.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:01863609429bdde21eb3473e4cf95739e54347819109a4d54c5a249f872b4fc6 +size 623 diff --git a/checkpoint-21/trainer_state.json b/checkpoint-21/trainer_state.json new file mode 100644 index 0000000000000000000000000000000000000000..363b9c25a091926380ff95b9dace2ac0e959e99b --- /dev/null +++ b/checkpoint-21/trainer_state.json @@ -0,0 +1,40 @@ +{ + "best_metric": 0.0, + "best_model_checkpoint": "wav2vec2-base-is_vinyl_scratched_or_not/checkpoint-21", + "epoch": 0.9767441860465116, + "global_step": 21, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.05, + "learning_rate": 1.4285714285714286e-06, + "loss": 0.6901, + "step": 1 + }, + { + "epoch": 0.98, + "learning_rate": 3e-05, + "loss": 0.6671, + "step": 21 + }, + { + "epoch": 0.98, + "eval_F1": 0.0, + "eval_Precision": 0.0, + "eval_Recall": 0.0, + "eval_accuracy": 0.6559766763848397, + "eval_loss": 0.6235492825508118, + "eval_runtime": 1064.9816, + "eval_samples_per_second": 0.644, + "eval_steps_per_second": 0.021, + "step": 21 + } + ], + "max_steps": 210, + "num_train_epochs": 10, + "total_flos": 1.3691485884384e+17, + "trial_name": null, + "trial_params": null +} diff --git a/checkpoint-21/training_args.bin b/checkpoint-21/training_args.bin new file mode 100644 index 0000000000000000000000000000000000000000..44d4af4d6c45e0f8ad8da6879b0acf5f4add727f --- /dev/null +++ b/checkpoint-21/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4521181070c5206d39fa1359aeb62443676d9d133bf030a8d47fe7908ee32f60 +size 3503 diff --git a/checkpoint-210/config.json b/checkpoint-210/config.json new file mode 100644 index 0000000000000000000000000000000000000000..5895b668646a423372f040c92bb60d42a5449727 --- /dev/null +++ b/checkpoint-210/config.json @@ -0,0 +1,125 @@ +{ + "_name_or_path": "facebook/wav2vec2-base", + "activation_dropout": 0.0, + "adapter_kernel_size": 3, + "adapter_stride": 2, + "add_adapter": false, + "apply_spec_augment": true, + "architectures": [ + "Wav2Vec2ForSequenceClassification" + ], + "attention_dropout": 0.1, + "bos_token_id": 1, + "classifier_proj_size": 256, + "codevector_dim": 256, + "contrastive_logits_temperature": 0.1, + "conv_bias": false, + "conv_dim": [ + 512, + 512, + 512, + 512, + 512, + 512, + 512 + ], + "conv_kernel": [ + 10, + 3, + 3, + 3, + 3, + 2, + 2 + ], + "conv_stride": [ + 5, + 2, + 2, + 2, + 2, + 2, + 2 + ], + "ctc_loss_reduction": "sum", + "ctc_zero_infinity": false, + "diversity_loss_weight": 0.1, + "do_stable_layer_norm": false, + "eos_token_id": 2, + "feat_extract_activation": "gelu", + "feat_extract_norm": "group", + "feat_proj_dropout": 0.1, + "feat_quantizer_dropout": 0.0, + "final_dropout": 0.0, + "freeze_feat_extract_train": true, + "hidden_act": "gelu", + "hidden_dropout": 0.1, + "hidden_size": 768, + "id2label": { + "0": 0, + "1": 1 + }, + "initializer_range": 0.02, + "intermediate_size": 3072, + "label2id": { + "0": "0", + "1": "1" + }, + "layer_norm_eps": 1e-05, + "layerdrop": 0.0, + "mask_channel_length": 10, + "mask_channel_min_space": 1, + "mask_channel_other": 0.0, + "mask_channel_prob": 0.0, + "mask_channel_selection": "static", + "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_min_space": 1, + "mask_time_other": 0.0, + "mask_time_prob": 0.05, + "mask_time_selection": "static", + "model_type": "wav2vec2", + "no_mask_channel_overlap": false, + "no_mask_time_overlap": false, + "num_adapter_layers": 3, + "num_attention_heads": 12, + "num_codevector_groups": 2, + "num_codevectors_per_group": 320, + "num_conv_pos_embedding_groups": 16, + "num_conv_pos_embeddings": 128, + "num_feat_extract_layers": 7, + "num_hidden_layers": 12, + "num_negatives": 100, + "output_hidden_size": 768, + "pad_token_id": 0, + "proj_codevector_dim": 256, + "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": "4.26.0", + "use_weighted_layer_sum": false, + "vocab_size": 32, + "xvector_output_dim": 512 +} diff --git a/checkpoint-210/optimizer.pt b/checkpoint-210/optimizer.pt new file mode 100644 index 0000000000000000000000000000000000000000..da79990c251b54b49c1b5fa7d6b001abc21ca3fa --- /dev/null +++ b/checkpoint-210/optimizer.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fa7bfe072cd4d501bcf35334dedbce73c21990c9d3762c3c22ff0e833bf7b7cf +size 756671269 diff --git a/checkpoint-210/preprocessor_config.json b/checkpoint-210/preprocessor_config.json new file mode 100644 index 0000000000000000000000000000000000000000..a0b7227fc1d916e469b14f6c154ad6dfea1e6891 --- /dev/null +++ b/checkpoint-210/preprocessor_config.json @@ -0,0 +1,9 @@ +{ + "do_normalize": true, + "feature_extractor_type": "Wav2Vec2FeatureExtractor", + "feature_size": 1, + "padding_side": "right", + "padding_value": 0.0, + "return_attention_mask": false, + "sampling_rate": 16000 +} diff --git a/checkpoint-210/pytorch_model.bin b/checkpoint-210/pytorch_model.bin new file mode 100644 index 0000000000000000000000000000000000000000..c5c9d4020fa5d7af5df8d3234e924c4c907f7d38 --- /dev/null +++ b/checkpoint-210/pytorch_model.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:63bf4cd2d849a6267b9d0e24d3c3a05f20a5dfedbb58135701615cef6761cd88 +size 378345555 diff --git a/checkpoint-210/rng_state.pth b/checkpoint-210/rng_state.pth new file mode 100644 index 0000000000000000000000000000000000000000..895091512010cef60cdbb7339ec0a5cfa455eb38 --- /dev/null +++ b/checkpoint-210/rng_state.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:85cad62ef11f52bf98a70abd7f8acfd2eb03078ce3aedb5551ea84cc36bfa13b +size 13547 diff --git a/checkpoint-210/scheduler.pt b/checkpoint-210/scheduler.pt new file mode 100644 index 0000000000000000000000000000000000000000..96fa3ae6edc86627c5e1c76cd13fb5ea420c3839 --- /dev/null +++ b/checkpoint-210/scheduler.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:049f96322d11654bf522b339cfc10eec25479db7d28d0c95cec8a05487bff74d +size 623 diff --git a/checkpoint-210/trainer_state.json b/checkpoint-210/trainer_state.json new file mode 100644 index 0000000000000000000000000000000000000000..69c114512ca48ac969e26048fff7f2c25845e582 --- /dev/null +++ b/checkpoint-210/trainer_state.json @@ -0,0 +1,202 @@ +{ + "best_metric": 0.9637526652452025, + "best_model_checkpoint": "wav2vec2-base-is_vinyl_scratched_or_not/checkpoint-168", + "epoch": 9.976744186046512, + "global_step": 210, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.05, + "learning_rate": 1.4285714285714286e-06, + "loss": 0.6901, + "step": 1 + }, + { + "epoch": 0.98, + "learning_rate": 3e-05, + "loss": 0.6671, + "step": 21 + }, + { + "epoch": 0.98, + "eval_F1": 0.0, + "eval_Precision": 0.0, + "eval_Recall": 0.0, + "eval_accuracy": 0.6559766763848397, + "eval_loss": 0.6235492825508118, + "eval_runtime": 1064.9816, + "eval_samples_per_second": 0.644, + "eval_steps_per_second": 0.021, + "step": 21 + }, + { + "epoch": 1.98, + "learning_rate": 2.6666666666666667e-05, + "loss": 0.4954, + "step": 42 + }, + { + "epoch": 1.98, + "eval_F1": 0.9095022624434389, + "eval_Precision": 0.9757281553398058, + "eval_Recall": 0.8516949152542372, + "eval_accuracy": 0.9416909620991254, + "eval_loss": 0.28244873881340027, + "eval_runtime": 1042.1602, + "eval_samples_per_second": 0.658, + "eval_steps_per_second": 0.021, + "step": 42 + }, + { + "epoch": 2.98, + "learning_rate": 2.3333333333333336e-05, + "loss": 0.2406, + "step": 63 + }, + { + "epoch": 2.98, + "eval_F1": 0.9336283185840708, + "eval_Precision": 0.9768518518518519, + "eval_Recall": 0.8940677966101694, + "eval_accuracy": 0.956268221574344, + "eval_loss": 0.17554502189159393, + "eval_runtime": 1040.1013, + "eval_samples_per_second": 0.66, + "eval_steps_per_second": 0.021, + "step": 63 + }, + { + "epoch": 3.98, + "learning_rate": 1.9999999999999998e-05, + "loss": 0.169, + "step": 84 + }, + { + "epoch": 3.98, + "eval_F1": 0.9385964912280702, + "eval_Precision": 0.9727272727272728, + "eval_Recall": 0.9067796610169492, + "eval_accuracy": 0.9591836734693877, + "eval_loss": 0.1545204222202301, + "eval_runtime": 1051.4657, + "eval_samples_per_second": 0.652, + "eval_steps_per_second": 0.021, + "step": 84 + }, + { + "epoch": 4.98, + "learning_rate": 1.6666666666666667e-05, + "loss": 0.1287, + "step": 105 + }, + { + "epoch": 4.98, + "eval_F1": 0.9406593406593408, + "eval_Precision": 0.9771689497716894, + "eval_Recall": 0.9067796610169492, + "eval_accuracy": 0.9606413994169096, + "eval_loss": 0.1248895600438118, + "eval_runtime": 1050.486, + "eval_samples_per_second": 0.653, + "eval_steps_per_second": 0.021, + "step": 105 + }, + { + "epoch": 5.98, + "learning_rate": 1.3333333333333333e-05, + "loss": 0.1102, + "step": 126 + }, + { + "epoch": 5.98, + "eval_F1": 0.9594882729211086, + "eval_Precision": 0.9656652360515021, + "eval_Recall": 0.9533898305084746, + "eval_accuracy": 0.9723032069970845, + "eval_loss": 0.11587227135896683, + "eval_runtime": 1049.2012, + "eval_samples_per_second": 0.654, + "eval_steps_per_second": 0.021, + "step": 126 + }, + { + "epoch": 6.98, + "learning_rate": 9.999999999999999e-06, + "loss": 0.0923, + "step": 147 + }, + { + "epoch": 6.98, + "eval_F1": 0.9515789473684211, + "eval_Precision": 0.9456066945606695, + "eval_Recall": 0.9576271186440678, + "eval_accuracy": 0.9664723032069971, + "eval_loss": 0.10729651153087616, + "eval_runtime": 1059.4975, + "eval_samples_per_second": 0.647, + "eval_steps_per_second": 0.021, + "step": 147 + }, + { + "epoch": 7.98, + "learning_rate": 6.666666666666667e-06, + "loss": 0.0877, + "step": 168 + }, + { + "epoch": 7.98, + "eval_F1": 0.9637526652452025, + "eval_Precision": 0.9699570815450643, + "eval_Recall": 0.9576271186440678, + "eval_accuracy": 0.9752186588921283, + "eval_loss": 0.10392837971448898, + "eval_runtime": 1046.4548, + "eval_samples_per_second": 0.656, + "eval_steps_per_second": 0.021, + "step": 168 + }, + { + "epoch": 8.98, + "learning_rate": 3.3333333333333333e-06, + "loss": 0.0807, + "step": 189 + }, + { + "epoch": 8.98, + "eval_F1": 0.9535864978902954, + "eval_Precision": 0.9495798319327731, + "eval_Recall": 0.9576271186440678, + "eval_accuracy": 0.967930029154519, + "eval_loss": 0.1087886393070221, + "eval_runtime": 1049.1807, + "eval_samples_per_second": 0.654, + "eval_steps_per_second": 0.021, + "step": 189 + }, + { + "epoch": 9.98, + "learning_rate": 0.0, + "loss": 0.0744, + "step": 210 + }, + { + "epoch": 9.98, + "eval_F1": 0.9637526652452025, + "eval_Precision": 0.9699570815450643, + "eval_Recall": 0.9576271186440678, + "eval_accuracy": 0.9752186588921283, + "eval_loss": 0.1040809378027916, + "eval_runtime": 1111.0181, + "eval_samples_per_second": 0.617, + "eval_steps_per_second": 0.02, + "step": 210 + } + ], + "max_steps": 210, + "num_train_epochs": 10, + "total_flos": 1.36645223454432e+18, + "trial_name": null, + "trial_params": null +} diff --git a/checkpoint-210/training_args.bin b/checkpoint-210/training_args.bin new file mode 100644 index 0000000000000000000000000000000000000000..44d4af4d6c45e0f8ad8da6879b0acf5f4add727f --- /dev/null +++ b/checkpoint-210/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4521181070c5206d39fa1359aeb62443676d9d133bf030a8d47fe7908ee32f60 +size 3503 diff --git a/checkpoint-42/config.json b/checkpoint-42/config.json new file mode 100644 index 0000000000000000000000000000000000000000..5895b668646a423372f040c92bb60d42a5449727 --- /dev/null +++ b/checkpoint-42/config.json @@ -0,0 +1,125 @@ +{ + "_name_or_path": "facebook/wav2vec2-base", + "activation_dropout": 0.0, + "adapter_kernel_size": 3, + "adapter_stride": 2, + "add_adapter": false, + "apply_spec_augment": true, + "architectures": [ + "Wav2Vec2ForSequenceClassification" + ], + "attention_dropout": 0.1, + "bos_token_id": 1, + "classifier_proj_size": 256, + "codevector_dim": 256, + "contrastive_logits_temperature": 0.1, + "conv_bias": false, + "conv_dim": [ + 512, + 512, + 512, + 512, + 512, + 512, + 512 + ], + "conv_kernel": [ + 10, + 3, + 3, + 3, + 3, + 2, + 2 + ], + "conv_stride": [ + 5, + 2, + 2, + 2, + 2, + 2, + 2 + ], + "ctc_loss_reduction": "sum", + "ctc_zero_infinity": false, + "diversity_loss_weight": 0.1, + "do_stable_layer_norm": false, + "eos_token_id": 2, + "feat_extract_activation": "gelu", + "feat_extract_norm": "group", + "feat_proj_dropout": 0.1, + "feat_quantizer_dropout": 0.0, + "final_dropout": 0.0, + "freeze_feat_extract_train": true, + "hidden_act": "gelu", + "hidden_dropout": 0.1, + "hidden_size": 768, + "id2label": { + "0": 0, + "1": 1 + }, + "initializer_range": 0.02, + "intermediate_size": 3072, + "label2id": { + "0": "0", + "1": "1" + }, + "layer_norm_eps": 1e-05, + "layerdrop": 0.0, + "mask_channel_length": 10, + "mask_channel_min_space": 1, + "mask_channel_other": 0.0, + "mask_channel_prob": 0.0, + "mask_channel_selection": "static", + "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_min_space": 1, + "mask_time_other": 0.0, + "mask_time_prob": 0.05, + "mask_time_selection": "static", + "model_type": "wav2vec2", + "no_mask_channel_overlap": false, + "no_mask_time_overlap": false, + "num_adapter_layers": 3, + "num_attention_heads": 12, + "num_codevector_groups": 2, + "num_codevectors_per_group": 320, + "num_conv_pos_embedding_groups": 16, + "num_conv_pos_embeddings": 128, + "num_feat_extract_layers": 7, + "num_hidden_layers": 12, + "num_negatives": 100, + "output_hidden_size": 768, + "pad_token_id": 0, + "proj_codevector_dim": 256, + "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": "4.26.0", + "use_weighted_layer_sum": false, + "vocab_size": 32, + "xvector_output_dim": 512 +} diff --git a/checkpoint-42/optimizer.pt b/checkpoint-42/optimizer.pt new file mode 100644 index 0000000000000000000000000000000000000000..a88f542ee007a4d3cfd500b5da026592afd8f8a4 --- /dev/null +++ b/checkpoint-42/optimizer.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:dcacc8337e0ea5b1fb3930e5175e8999a9417758af90d28f92252181faa0952c +size 756671269 diff --git a/checkpoint-42/preprocessor_config.json b/checkpoint-42/preprocessor_config.json new file mode 100644 index 0000000000000000000000000000000000000000..a0b7227fc1d916e469b14f6c154ad6dfea1e6891 --- /dev/null +++ b/checkpoint-42/preprocessor_config.json @@ -0,0 +1,9 @@ +{ + "do_normalize": true, + "feature_extractor_type": "Wav2Vec2FeatureExtractor", + "feature_size": 1, + "padding_side": "right", + "padding_value": 0.0, + "return_attention_mask": false, + "sampling_rate": 16000 +} diff --git a/checkpoint-42/pytorch_model.bin b/checkpoint-42/pytorch_model.bin new file mode 100644 index 0000000000000000000000000000000000000000..ade48cb3b831693a565eae77b7a8151dd37ac757 --- /dev/null +++ b/checkpoint-42/pytorch_model.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d0a5cab0208308c13ab2bc9b3b34bb652c346df1b3d717d035148d13b091c713 +size 378345555 diff --git a/checkpoint-42/rng_state.pth b/checkpoint-42/rng_state.pth new file mode 100644 index 0000000000000000000000000000000000000000..cf84c016575a1e8bbd41cadc5f8aac2e13bfabb8 --- /dev/null +++ b/checkpoint-42/rng_state.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f343ebff3b1341827f796c69f566e874e596fca6703c296c48a24a797c636c7f +size 13611 diff --git a/checkpoint-42/scheduler.pt b/checkpoint-42/scheduler.pt new file mode 100644 index 0000000000000000000000000000000000000000..3386d3eac768ff765e27fe170380192a401625ed --- /dev/null +++ b/checkpoint-42/scheduler.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1d0cbe7be967389135fc7a69463fb158ab3d25206a4409658e9dd078ff70ed38 +size 623 diff --git a/checkpoint-42/trainer_state.json b/checkpoint-42/trainer_state.json new file mode 100644 index 0000000000000000000000000000000000000000..e0f481ee3367a95a5d00b9dbd79cfc8df5ee4f7e --- /dev/null +++ b/checkpoint-42/trainer_state.json @@ -0,0 +1,58 @@ +{ + "best_metric": 0.9095022624434389, + "best_model_checkpoint": "wav2vec2-base-is_vinyl_scratched_or_not/checkpoint-42", + "epoch": 1.9767441860465116, + "global_step": 42, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.05, + "learning_rate": 1.4285714285714286e-06, + "loss": 0.6901, + "step": 1 + }, + { + "epoch": 0.98, + "learning_rate": 3e-05, + "loss": 0.6671, + "step": 21 + }, + { + "epoch": 0.98, + "eval_F1": 0.0, + "eval_Precision": 0.0, + "eval_Recall": 0.0, + "eval_accuracy": 0.6559766763848397, + "eval_loss": 0.6235492825508118, + "eval_runtime": 1064.9816, + "eval_samples_per_second": 0.644, + "eval_steps_per_second": 0.021, + "step": 21 + }, + { + "epoch": 1.98, + "learning_rate": 2.6666666666666667e-05, + "loss": 0.4954, + "step": 42 + }, + { + "epoch": 1.98, + "eval_F1": 0.9095022624434389, + "eval_Precision": 0.9757281553398058, + "eval_Recall": 0.8516949152542372, + "eval_accuracy": 0.9416909620991254, + "eval_loss": 0.28244873881340027, + "eval_runtime": 1042.1602, + "eval_samples_per_second": 0.658, + "eval_steps_per_second": 0.021, + "step": 42 + } + ], + "max_steps": 210, + "num_train_epochs": 10, + "total_flos": 2.7382971768768e+17, + "trial_name": null, + "trial_params": null +} diff --git a/checkpoint-42/training_args.bin b/checkpoint-42/training_args.bin new file mode 100644 index 0000000000000000000000000000000000000000..44d4af4d6c45e0f8ad8da6879b0acf5f4add727f --- /dev/null +++ b/checkpoint-42/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4521181070c5206d39fa1359aeb62443676d9d133bf030a8d47fe7908ee32f60 +size 3503 diff --git a/checkpoint-63/config.json b/checkpoint-63/config.json new file mode 100644 index 0000000000000000000000000000000000000000..5895b668646a423372f040c92bb60d42a5449727 --- /dev/null +++ b/checkpoint-63/config.json @@ -0,0 +1,125 @@ +{ + "_name_or_path": "facebook/wav2vec2-base", + "activation_dropout": 0.0, + "adapter_kernel_size": 3, + "adapter_stride": 2, + "add_adapter": false, + "apply_spec_augment": true, + "architectures": [ + "Wav2Vec2ForSequenceClassification" + ], + "attention_dropout": 0.1, + "bos_token_id": 1, + "classifier_proj_size": 256, + "codevector_dim": 256, + "contrastive_logits_temperature": 0.1, + "conv_bias": false, + "conv_dim": [ + 512, + 512, + 512, + 512, + 512, + 512, + 512 + ], + "conv_kernel": [ + 10, + 3, + 3, + 3, + 3, + 2, + 2 + ], + "conv_stride": [ + 5, + 2, + 2, + 2, + 2, + 2, + 2 + ], + "ctc_loss_reduction": "sum", + "ctc_zero_infinity": false, + "diversity_loss_weight": 0.1, + "do_stable_layer_norm": false, + "eos_token_id": 2, + "feat_extract_activation": "gelu", + "feat_extract_norm": "group", + "feat_proj_dropout": 0.1, + "feat_quantizer_dropout": 0.0, + "final_dropout": 0.0, + "freeze_feat_extract_train": true, + "hidden_act": "gelu", + "hidden_dropout": 0.1, + "hidden_size": 768, + "id2label": { + "0": 0, + "1": 1 + }, + "initializer_range": 0.02, + "intermediate_size": 3072, + "label2id": { + "0": "0", + "1": "1" + }, + "layer_norm_eps": 1e-05, + "layerdrop": 0.0, + "mask_channel_length": 10, + "mask_channel_min_space": 1, + "mask_channel_other": 0.0, + "mask_channel_prob": 0.0, + "mask_channel_selection": "static", + "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_min_space": 1, + "mask_time_other": 0.0, + "mask_time_prob": 0.05, + "mask_time_selection": "static", + "model_type": "wav2vec2", + "no_mask_channel_overlap": false, + "no_mask_time_overlap": false, + "num_adapter_layers": 3, + "num_attention_heads": 12, + "num_codevector_groups": 2, + "num_codevectors_per_group": 320, + "num_conv_pos_embedding_groups": 16, + "num_conv_pos_embeddings": 128, + "num_feat_extract_layers": 7, + "num_hidden_layers": 12, + "num_negatives": 100, + "output_hidden_size": 768, + "pad_token_id": 0, + "proj_codevector_dim": 256, + "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": "4.26.0", + "use_weighted_layer_sum": false, + "vocab_size": 32, + "xvector_output_dim": 512 +} diff --git a/checkpoint-63/optimizer.pt b/checkpoint-63/optimizer.pt new file mode 100644 index 0000000000000000000000000000000000000000..97c90184ba493acad8ab80ea96b871dd91290abe --- /dev/null +++ b/checkpoint-63/optimizer.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e288611b9cbc4ee0d24ea542d105bc8402465d36c862901b136a2ddc8922b8fe +size 756671269 diff --git a/checkpoint-63/preprocessor_config.json b/checkpoint-63/preprocessor_config.json new file mode 100644 index 0000000000000000000000000000000000000000..a0b7227fc1d916e469b14f6c154ad6dfea1e6891 --- /dev/null +++ b/checkpoint-63/preprocessor_config.json @@ -0,0 +1,9 @@ +{ + "do_normalize": true, + "feature_extractor_type": "Wav2Vec2FeatureExtractor", + "feature_size": 1, + "padding_side": "right", + "padding_value": 0.0, + "return_attention_mask": false, + "sampling_rate": 16000 +} diff --git a/checkpoint-63/pytorch_model.bin b/checkpoint-63/pytorch_model.bin new file mode 100644 index 0000000000000000000000000000000000000000..d573e6be87aab4921948bca379ed3a9a88951998 --- /dev/null +++ b/checkpoint-63/pytorch_model.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e3cb509572d40bc293a4f66213fc139811d7b5a7ef0e3fc4b5fc02b86b8b0085 +size 378345555 diff --git a/checkpoint-63/rng_state.pth b/checkpoint-63/rng_state.pth new file mode 100644 index 0000000000000000000000000000000000000000..38a47deab08cf1326e57e43aafa54525a74224e9 --- /dev/null +++ b/checkpoint-63/rng_state.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f93b1171d33339a4634159a2f818e7612416d525af518a0e871c42ca9428f1a5 +size 13611 diff --git a/checkpoint-63/scheduler.pt b/checkpoint-63/scheduler.pt new file mode 100644 index 0000000000000000000000000000000000000000..d2fb891bddfd1dc6faa8189a7c135cdf6935b837 --- /dev/null +++ b/checkpoint-63/scheduler.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8ea28cc8b11a5a41d268d99f63d3b63644dc52a4f6a53a6ca4fa67675a944e29 +size 623 diff --git a/checkpoint-63/trainer_state.json b/checkpoint-63/trainer_state.json new file mode 100644 index 0000000000000000000000000000000000000000..a421e3c6da0505129eb1d9aeb7b16df59341eb3e --- /dev/null +++ b/checkpoint-63/trainer_state.json @@ -0,0 +1,76 @@ +{ + "best_metric": 0.9336283185840708, + "best_model_checkpoint": "wav2vec2-base-is_vinyl_scratched_or_not/checkpoint-63", + "epoch": 2.9767441860465116, + "global_step": 63, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.05, + "learning_rate": 1.4285714285714286e-06, + "loss": 0.6901, + "step": 1 + }, + { + "epoch": 0.98, + "learning_rate": 3e-05, + "loss": 0.6671, + "step": 21 + }, + { + "epoch": 0.98, + "eval_F1": 0.0, + "eval_Precision": 0.0, + "eval_Recall": 0.0, + "eval_accuracy": 0.6559766763848397, + "eval_loss": 0.6235492825508118, + "eval_runtime": 1064.9816, + "eval_samples_per_second": 0.644, + "eval_steps_per_second": 0.021, + "step": 21 + }, + { + "epoch": 1.98, + "learning_rate": 2.6666666666666667e-05, + "loss": 0.4954, + "step": 42 + }, + { + "epoch": 1.98, + "eval_F1": 0.9095022624434389, + "eval_Precision": 0.9757281553398058, + "eval_Recall": 0.8516949152542372, + "eval_accuracy": 0.9416909620991254, + "eval_loss": 0.28244873881340027, + "eval_runtime": 1042.1602, + "eval_samples_per_second": 0.658, + "eval_steps_per_second": 0.021, + "step": 42 + }, + { + "epoch": 2.98, + "learning_rate": 2.3333333333333336e-05, + "loss": 0.2406, + "step": 63 + }, + { + "epoch": 2.98, + "eval_F1": 0.9336283185840708, + "eval_Precision": 0.9768518518518519, + "eval_Recall": 0.8940677966101694, + "eval_accuracy": 0.956268221574344, + "eval_loss": 0.17554502189159393, + "eval_runtime": 1040.1013, + "eval_samples_per_second": 0.66, + "eval_steps_per_second": 0.021, + "step": 63 + } + ], + "max_steps": 210, + "num_train_epochs": 10, + "total_flos": 4.1074457653152e+17, + "trial_name": null, + "trial_params": null +} diff --git a/checkpoint-63/training_args.bin b/checkpoint-63/training_args.bin new file mode 100644 index 0000000000000000000000000000000000000000..44d4af4d6c45e0f8ad8da6879b0acf5f4add727f --- /dev/null +++ b/checkpoint-63/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4521181070c5206d39fa1359aeb62443676d9d133bf030a8d47fe7908ee32f60 +size 3503 diff --git a/checkpoint-84/config.json b/checkpoint-84/config.json new file mode 100644 index 0000000000000000000000000000000000000000..5895b668646a423372f040c92bb60d42a5449727 --- /dev/null +++ b/checkpoint-84/config.json @@ -0,0 +1,125 @@ +{ + "_name_or_path": "facebook/wav2vec2-base", + "activation_dropout": 0.0, + "adapter_kernel_size": 3, + "adapter_stride": 2, + "add_adapter": false, + "apply_spec_augment": true, + "architectures": [ + "Wav2Vec2ForSequenceClassification" + ], + "attention_dropout": 0.1, + "bos_token_id": 1, + "classifier_proj_size": 256, + "codevector_dim": 256, + "contrastive_logits_temperature": 0.1, + "conv_bias": false, + "conv_dim": [ + 512, + 512, + 512, + 512, + 512, + 512, + 512 + ], + "conv_kernel": [ + 10, + 3, + 3, + 3, + 3, + 2, + 2 + ], + "conv_stride": [ + 5, + 2, + 2, + 2, + 2, + 2, + 2 + ], + "ctc_loss_reduction": "sum", + "ctc_zero_infinity": false, + "diversity_loss_weight": 0.1, + "do_stable_layer_norm": false, + "eos_token_id": 2, + "feat_extract_activation": "gelu", + "feat_extract_norm": "group", + "feat_proj_dropout": 0.1, + "feat_quantizer_dropout": 0.0, + "final_dropout": 0.0, + "freeze_feat_extract_train": true, + "hidden_act": "gelu", + "hidden_dropout": 0.1, + "hidden_size": 768, + "id2label": { + "0": 0, + "1": 1 + }, + "initializer_range": 0.02, + "intermediate_size": 3072, + "label2id": { + "0": "0", + "1": "1" + }, + "layer_norm_eps": 1e-05, + "layerdrop": 0.0, + "mask_channel_length": 10, + "mask_channel_min_space": 1, + "mask_channel_other": 0.0, + "mask_channel_prob": 0.0, + "mask_channel_selection": "static", + "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_min_space": 1, + "mask_time_other": 0.0, + "mask_time_prob": 0.05, + "mask_time_selection": "static", + "model_type": "wav2vec2", + "no_mask_channel_overlap": false, + "no_mask_time_overlap": false, + "num_adapter_layers": 3, + "num_attention_heads": 12, + "num_codevector_groups": 2, + "num_codevectors_per_group": 320, + "num_conv_pos_embedding_groups": 16, + "num_conv_pos_embeddings": 128, + "num_feat_extract_layers": 7, + "num_hidden_layers": 12, + "num_negatives": 100, + "output_hidden_size": 768, + "pad_token_id": 0, + "proj_codevector_dim": 256, + "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": "4.26.0", + "use_weighted_layer_sum": false, + "vocab_size": 32, + "xvector_output_dim": 512 +} diff --git a/checkpoint-84/optimizer.pt b/checkpoint-84/optimizer.pt new file mode 100644 index 0000000000000000000000000000000000000000..43d1d71f1a6ab576fa5807590f4da36970b6207f --- /dev/null +++ b/checkpoint-84/optimizer.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ecc569dbac6da16d22017039906df74e61056fd936b94dac3237792f112a069f +size 756671269 diff --git a/checkpoint-84/preprocessor_config.json b/checkpoint-84/preprocessor_config.json new file mode 100644 index 0000000000000000000000000000000000000000..a0b7227fc1d916e469b14f6c154ad6dfea1e6891 --- /dev/null +++ b/checkpoint-84/preprocessor_config.json @@ -0,0 +1,9 @@ +{ + "do_normalize": true, + "feature_extractor_type": "Wav2Vec2FeatureExtractor", + "feature_size": 1, + "padding_side": "right", + "padding_value": 0.0, + "return_attention_mask": false, + "sampling_rate": 16000 +} diff --git a/checkpoint-84/pytorch_model.bin b/checkpoint-84/pytorch_model.bin new file mode 100644 index 0000000000000000000000000000000000000000..479ec0a7653c1ccaa68fafe48b5bdcc0b39b19ed --- /dev/null +++ b/checkpoint-84/pytorch_model.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9ba55ac4d586182c03a5d022481cf10ae6bf9773c0de5d02262c34ef29aeec4a +size 378345555 diff --git a/checkpoint-84/rng_state.pth b/checkpoint-84/rng_state.pth new file mode 100644 index 0000000000000000000000000000000000000000..44895fc0e201bea36b0023de323510bb7db76edf --- /dev/null +++ b/checkpoint-84/rng_state.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e18266cb9f7da25f5d884957cf904db9fca6853e7fb32b639d5299815842ed49 +size 13547 diff --git a/checkpoint-84/scheduler.pt b/checkpoint-84/scheduler.pt new file mode 100644 index 0000000000000000000000000000000000000000..39cca5bd6e5676ca72ef6f801cba8332c6ee1aaa --- /dev/null +++ b/checkpoint-84/scheduler.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:15904d72bb4864895f6558e60b1529e2ca14de12f06fb933f9ad322e766b8af3 +size 623 diff --git a/checkpoint-84/trainer_state.json b/checkpoint-84/trainer_state.json new file mode 100644 index 0000000000000000000000000000000000000000..2224a55d26e91c090a2023b702ff278072c71aec --- /dev/null +++ b/checkpoint-84/trainer_state.json @@ -0,0 +1,94 @@ +{ + "best_metric": 0.9385964912280702, + "best_model_checkpoint": "wav2vec2-base-is_vinyl_scratched_or_not/checkpoint-84", + "epoch": 3.9767441860465116, + "global_step": 84, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.05, + "learning_rate": 1.4285714285714286e-06, + "loss": 0.6901, + "step": 1 + }, + { + "epoch": 0.98, + "learning_rate": 3e-05, + "loss": 0.6671, + "step": 21 + }, + { + "epoch": 0.98, + "eval_F1": 0.0, + "eval_Precision": 0.0, + "eval_Recall": 0.0, + "eval_accuracy": 0.6559766763848397, + "eval_loss": 0.6235492825508118, + "eval_runtime": 1064.9816, + "eval_samples_per_second": 0.644, + "eval_steps_per_second": 0.021, + "step": 21 + }, + { + "epoch": 1.98, + "learning_rate": 2.6666666666666667e-05, + "loss": 0.4954, + "step": 42 + }, + { + "epoch": 1.98, + "eval_F1": 0.9095022624434389, + "eval_Precision": 0.9757281553398058, + "eval_Recall": 0.8516949152542372, + "eval_accuracy": 0.9416909620991254, + "eval_loss": 0.28244873881340027, + "eval_runtime": 1042.1602, + "eval_samples_per_second": 0.658, + "eval_steps_per_second": 0.021, + "step": 42 + }, + { + "epoch": 2.98, + "learning_rate": 2.3333333333333336e-05, + "loss": 0.2406, + "step": 63 + }, + { + "epoch": 2.98, + "eval_F1": 0.9336283185840708, + "eval_Precision": 0.9768518518518519, + "eval_Recall": 0.8940677966101694, + "eval_accuracy": 0.956268221574344, + "eval_loss": 0.17554502189159393, + "eval_runtime": 1040.1013, + "eval_samples_per_second": 0.66, + "eval_steps_per_second": 0.021, + "step": 63 + }, + { + "epoch": 3.98, + "learning_rate": 1.9999999999999998e-05, + "loss": 0.169, + "step": 84 + }, + { + "epoch": 3.98, + "eval_F1": 0.9385964912280702, + "eval_Precision": 0.9727272727272728, + "eval_Recall": 0.9067796610169492, + "eval_accuracy": 0.9591836734693877, + "eval_loss": 0.1545204222202301, + "eval_runtime": 1051.4657, + "eval_samples_per_second": 0.652, + "eval_steps_per_second": 0.021, + "step": 84 + } + ], + "max_steps": 210, + "num_train_epochs": 10, + "total_flos": 5.4765943537536e+17, + "trial_name": null, + "trial_params": null +} diff --git a/checkpoint-84/training_args.bin b/checkpoint-84/training_args.bin new file mode 100644 index 0000000000000000000000000000000000000000..44d4af4d6c45e0f8ad8da6879b0acf5f4add727f --- /dev/null +++ b/checkpoint-84/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4521181070c5206d39fa1359aeb62443676d9d133bf030a8d47fe7908ee32f60 +size 3503 diff --git a/config.json b/config.json new file mode 100644 index 0000000000000000000000000000000000000000..5895b668646a423372f040c92bb60d42a5449727 --- /dev/null +++ b/config.json @@ -0,0 +1,125 @@ +{ + "_name_or_path": "facebook/wav2vec2-base", + "activation_dropout": 0.0, + "adapter_kernel_size": 3, + "adapter_stride": 2, + "add_adapter": false, + "apply_spec_augment": true, + "architectures": [ + "Wav2Vec2ForSequenceClassification" + ], + "attention_dropout": 0.1, + "bos_token_id": 1, + "classifier_proj_size": 256, + "codevector_dim": 256, + "contrastive_logits_temperature": 0.1, + "conv_bias": false, + "conv_dim": [ + 512, + 512, + 512, + 512, + 512, + 512, + 512 + ], + "conv_kernel": [ + 10, + 3, + 3, + 3, + 3, + 2, + 2 + ], + "conv_stride": [ + 5, + 2, + 2, + 2, + 2, + 2, + 2 + ], + "ctc_loss_reduction": "sum", + "ctc_zero_infinity": false, + "diversity_loss_weight": 0.1, + "do_stable_layer_norm": false, + "eos_token_id": 2, + "feat_extract_activation": "gelu", + "feat_extract_norm": "group", + "feat_proj_dropout": 0.1, + "feat_quantizer_dropout": 0.0, + "final_dropout": 0.0, + "freeze_feat_extract_train": true, + "hidden_act": "gelu", + "hidden_dropout": 0.1, + "hidden_size": 768, + "id2label": { + "0": 0, + "1": 1 + }, + "initializer_range": 0.02, + "intermediate_size": 3072, + "label2id": { + "0": "0", + "1": "1" + }, + "layer_norm_eps": 1e-05, + "layerdrop": 0.0, + "mask_channel_length": 10, + "mask_channel_min_space": 1, + "mask_channel_other": 0.0, + "mask_channel_prob": 0.0, + "mask_channel_selection": "static", + "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_min_space": 1, + "mask_time_other": 0.0, + "mask_time_prob": 0.05, + "mask_time_selection": "static", + "model_type": "wav2vec2", + "no_mask_channel_overlap": false, + "no_mask_time_overlap": false, + "num_adapter_layers": 3, + "num_attention_heads": 12, + "num_codevector_groups": 2, + "num_codevectors_per_group": 320, + "num_conv_pos_embedding_groups": 16, + "num_conv_pos_embeddings": 128, + "num_feat_extract_layers": 7, + "num_hidden_layers": 12, + "num_negatives": 100, + "output_hidden_size": 768, + "pad_token_id": 0, + "proj_codevector_dim": 256, + "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": "4.26.0", + "use_weighted_layer_sum": false, + "vocab_size": 32, + "xvector_output_dim": 512 +} diff --git a/preprocessor_config.json b/preprocessor_config.json new file mode 100644 index 0000000000000000000000000000000000000000..a0b7227fc1d916e469b14f6c154ad6dfea1e6891 --- /dev/null +++ b/preprocessor_config.json @@ -0,0 +1,9 @@ +{ + "do_normalize": true, + "feature_extractor_type": "Wav2Vec2FeatureExtractor", + "feature_size": 1, + "padding_side": "right", + "padding_value": 0.0, + "return_attention_mask": false, + "sampling_rate": 16000 +} diff --git a/pytorch_model.bin b/pytorch_model.bin new file mode 100644 index 0000000000000000000000000000000000000000..529b102bf9e63664b07f7465cca4bf33eee5c841 --- /dev/null +++ b/pytorch_model.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2d80feebf75b9add887b3409fea764a26b226005fc477c13b47098cca4579b04 +size 378345555 diff --git a/training_args.bin b/training_args.bin new file mode 100644 index 0000000000000000000000000000000000000000..44d4af4d6c45e0f8ad8da6879b0acf5f4add727f --- /dev/null +++ b/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4521181070c5206d39fa1359aeb62443676d9d133bf030a8d47fe7908ee32f60 +size 3503