diff --git a/README.md b/README.md index 5d63f7e1d519ad5233855bfe9e5862b31c9bbfb2..17c49aa30b6f8e1eab777582762ed58716f2de03 100644 --- a/README.md +++ b/README.md @@ -6,9 +6,36 @@ tags: - generated_from_trainer datasets: - audiofolder +metrics: +- accuracy +- f1 +- precision +- recall model-index: - name: distilhubert-finetuned-cry-detector - results: [] + results: + - task: + name: Audio Classification + type: audio-classification + dataset: + name: audiofolder + type: audiofolder + config: default + split: train + args: default + metrics: + - name: Accuracy + type: accuracy + value: 0.9852941176470589 + - name: F1 + type: f1 + value: 0.9853150765112866 + - name: Precision + type: precision + value: 0.9853868369053048 + - name: Recall + type: recall + value: 0.9852941176470589 --- # distilhubert-finetuned-cry-detector This model is a fine-tuned version of [ntu-spml/distilhubert](https://huggingface.co/ntu-spml/distilhubert) on the audiofolder dataset. +It achieves the following results on the evaluation set: +- Loss: 0.0332 +- Accuracy: 0.9853 +- F1: 0.9853 +- Precision: 0.9854 +- Recall: 0.9853 ## Model description @@ -44,13 +77,22 @@ The following hyperparameters were used during training: - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: cosine - lr_scheduler_warmup_ratio: 0.001 -- num_epochs: 1 +- num_epochs: 10 ### Training results | Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 | Precision | Recall | |:-------------:|:------:|:----:|:---------------:|:--------:|:------:|:---------:|:------:| -| No log | 0.9412 | 12 | 0.3379 | 0.9559 | 0.9555 | 0.9564 | 0.9559 | +| No log | 0.9412 | 12 | 0.1931 | 0.9363 | 0.9365 | 0.9372 | 0.9363 | +| No log | 1.9608 | 25 | 0.0950 | 0.9706 | 0.9704 | 0.9710 | 0.9706 | +| No log | 2.9804 | 38 | 0.0611 | 0.9804 | 0.9804 | 0.9804 | 0.9804 | +| No log | 4.0 | 51 | 0.0492 | 0.9853 | 0.9853 | 0.9853 | 0.9853 | +| No log | 4.9412 | 63 | 0.0588 | 0.9804 | 0.9805 | 0.9814 | 0.9804 | +| No log | 5.9608 | 76 | 0.0368 | 0.9853 | 0.9853 | 0.9854 | 0.9853 | +| No log | 6.9804 | 89 | 0.0382 | 0.9902 | 0.9902 | 0.9903 | 0.9902 | +| No log | 8.0 | 102 | 0.0318 | 0.9951 | 0.9951 | 0.9951 | 0.9951 | +| No log | 8.9412 | 114 | 0.0331 | 0.9853 | 0.9853 | 0.9854 | 0.9853 | +| No log | 9.4118 | 120 | 0.0332 | 0.9853 | 0.9853 | 0.9854 | 0.9853 | ### Framework versions diff --git a/checkpoint-102/config.json b/checkpoint-102/config.json new file mode 100644 index 0000000000000000000000000000000000000000..4550a22abfa803293b5dc6b2633d0201ac2ff44d --- /dev/null +++ b/checkpoint-102/config.json @@ -0,0 +1,80 @@ +{ + "_name_or_path": "ntu-spml/distilhubert", + "activation_dropout": 0.1, + "apply_spec_augment": false, + "architectures": [ + "HubertForSequenceClassification" + ], + "attention_dropout": 0.1, + "bos_token_id": 1, + "classifier_proj_size": 256, + "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, + "do_stable_layer_norm": false, + "eos_token_id": 2, + "feat_extract_activation": "gelu", + "feat_extract_norm": "group", + "feat_proj_dropout": 0.0, + "feat_proj_layer_norm": false, + "final_dropout": 0.0, + "hidden_act": "gelu", + "hidden_dropout": 0.1, + "hidden_size": 768, + "id2label": { + "0": "crying", + "1": "no_crying" + }, + "initializer_range": 0.02, + "intermediate_size": 3072, + "label2id": { + "crying": "0", + "no_crying": "1" + }, + "layer_norm_eps": 1e-05, + "layerdrop": 0.0, + "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": 0.05, + "model_type": "hubert", + "num_attention_heads": 12, + "num_conv_pos_embedding_groups": 16, + "num_conv_pos_embeddings": 128, + "num_feat_extract_layers": 7, + "num_hidden_layers": 2, + "pad_token_id": 0, + "torch_dtype": "float32", + "transformers_version": "4.44.2", + "use_weighted_layer_sum": false, + "vocab_size": 32 +} diff --git a/checkpoint-102/model.safetensors b/checkpoint-102/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..b63e1b217d7bda6f42a6ac5fe5cc72b618e9bffe --- /dev/null +++ b/checkpoint-102/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:098c90d2cba195ac20033090c5954f7ef596abcdeebba724cb1365151033b05e +size 94763496 diff --git a/checkpoint-102/optimizer.pt b/checkpoint-102/optimizer.pt new file mode 100644 index 0000000000000000000000000000000000000000..dfad6acafb0bbfc8b8aa2e70dac6348e8f7730c1 --- /dev/null +++ b/checkpoint-102/optimizer.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6038d1ad5c694fa39b163e31b2ae8b290c47bf72179e6f7edebc2516088a6742 +size 189552570 diff --git a/checkpoint-102/preprocessor_config.json b/checkpoint-102/preprocessor_config.json new file mode 100644 index 0000000000000000000000000000000000000000..8ad80e2bc82023595cde1d13979733ba2dca60cc --- /dev/null +++ b/checkpoint-102/preprocessor_config.json @@ -0,0 +1,9 @@ +{ + "do_normalize": false, + "feature_extractor_type": "Wav2Vec2FeatureExtractor", + "feature_size": 1, + "padding_side": "right", + "padding_value": 0, + "return_attention_mask": false, + "sampling_rate": 16000 +} diff --git a/checkpoint-102/rng_state.pth b/checkpoint-102/rng_state.pth new file mode 100644 index 0000000000000000000000000000000000000000..b21c82b54d06f746e6decced536485e144604775 --- /dev/null +++ b/checkpoint-102/rng_state.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2e516bd48eedd1efdcdab1bc63eb245ee5f25372c9b0345efbc5ed2cd8b1b542 +size 14308 diff --git a/checkpoint-102/scheduler.pt b/checkpoint-102/scheduler.pt new file mode 100644 index 0000000000000000000000000000000000000000..6ffe09825ad80fe346a1160ac6328d5aac8f835b --- /dev/null +++ b/checkpoint-102/scheduler.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2b0f7d0de8839ca69a8acf7eae5d4a6ab5235d97ffbdfee2586222c3a1d5f473 +size 1064 diff --git a/checkpoint-102/trainer_state.json b/checkpoint-102/trainer_state.json new file mode 100644 index 0000000000000000000000000000000000000000..f299bdcae4764b760719f02bfc7fe5a6b67f2c93 --- /dev/null +++ b/checkpoint-102/trainer_state.json @@ -0,0 +1,138 @@ +{ + "best_metric": 0.9950980392156863, + "best_model_checkpoint": "distilhubert-finetuned-cry-detector/checkpoint-102", + "epoch": 8.0, + "eval_steps": 500, + "global_step": 102, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.9411764705882353, + "eval_accuracy": 0.9362745098039216, + "eval_f1": 0.9365356058160116, + "eval_loss": 0.19313116371631622, + "eval_precision": 0.9372234219385678, + "eval_recall": 0.9362745098039216, + "eval_runtime": 3.0986, + "eval_samples_per_second": 65.836, + "eval_steps_per_second": 8.391, + "step": 12 + }, + { + "epoch": 1.9607843137254903, + "eval_accuracy": 0.9705882352941176, + "eval_f1": 0.9704025549613785, + "eval_loss": 0.09495726972818375, + "eval_precision": 0.971039759187257, + "eval_recall": 0.9705882352941176, + "eval_runtime": 3.1171, + "eval_samples_per_second": 65.446, + "eval_steps_per_second": 8.341, + "step": 25 + }, + { + "epoch": 2.980392156862745, + "eval_accuracy": 0.9803921568627451, + "eval_f1": 0.9803921568627451, + "eval_loss": 0.06114682927727699, + "eval_precision": 0.9803921568627451, + "eval_recall": 0.9803921568627451, + "eval_runtime": 3.2163, + "eval_samples_per_second": 63.428, + "eval_steps_per_second": 8.084, + "step": 38 + }, + { + "epoch": 4.0, + "eval_accuracy": 0.9852941176470589, + "eval_f1": 0.9852722698274712, + "eval_loss": 0.049169763922691345, + "eval_precision": 0.9853018065952722, + "eval_recall": 0.9852941176470589, + "eval_runtime": 3.1223, + "eval_samples_per_second": 65.336, + "eval_steps_per_second": 8.327, + "step": 51 + }, + { + "epoch": 4.9411764705882355, + "eval_accuracy": 0.9803921568627451, + "eval_f1": 0.9804969685242518, + "eval_loss": 0.05879249796271324, + "eval_precision": 0.9813976872800402, + "eval_recall": 0.9803921568627451, + "eval_runtime": 3.1449, + "eval_samples_per_second": 64.868, + "eval_steps_per_second": 8.267, + "step": 63 + }, + { + "epoch": 5.96078431372549, + "eval_accuracy": 0.9852941176470589, + "eval_f1": 0.9853150765112866, + "eval_loss": 0.03680075705051422, + "eval_precision": 0.9853868369053048, + "eval_recall": 0.9852941176470589, + "eval_runtime": 3.1698, + "eval_samples_per_second": 64.358, + "eval_steps_per_second": 8.203, + "step": 76 + }, + { + "epoch": 6.980392156862745, + "eval_accuracy": 0.9901960784313726, + "eval_f1": 0.9901663478316141, + "eval_loss": 0.0382298119366169, + "eval_precision": 0.9903446226975638, + "eval_recall": 0.9901960784313726, + "eval_runtime": 3.1376, + "eval_samples_per_second": 65.019, + "eval_steps_per_second": 8.287, + "step": 89 + }, + { + "epoch": 8.0, + "eval_accuracy": 0.9950980392156863, + "eval_f1": 0.995090756609157, + "eval_loss": 0.03175761550664902, + "eval_precision": 0.995135458763658, + "eval_recall": 0.9950980392156863, + "eval_runtime": 2.949, + "eval_samples_per_second": 69.175, + "eval_steps_per_second": 8.816, + "step": 102 + } + ], + "logging_steps": 500, + "max_steps": 120, + "num_input_tokens_seen": 0, + "num_train_epochs": 10, + "save_steps": 500, + "stateful_callbacks": { + "EarlyStoppingCallback": { + "args": { + "early_stopping_patience": 3, + "early_stopping_threshold": 0.0 + }, + "attributes": { + "early_stopping_patience_counter": 0 + } + }, + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": false + }, + "attributes": {} + } + }, + "total_flos": 1.052064945063383e+17, + "train_batch_size": 8, + "trial_name": null, + "trial_params": null +} diff --git a/checkpoint-102/training_args.bin b/checkpoint-102/training_args.bin new file mode 100644 index 0000000000000000000000000000000000000000..ff5c682679da830384645399a104cd532805ce20 --- /dev/null +++ b/checkpoint-102/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3b77b2a33ca88330b7909aacab81332695e9c9dcf797e7a7757ca31ffcde6f31 +size 5240 diff --git a/checkpoint-114/config.json b/checkpoint-114/config.json new file mode 100644 index 0000000000000000000000000000000000000000..4550a22abfa803293b5dc6b2633d0201ac2ff44d --- /dev/null +++ b/checkpoint-114/config.json @@ -0,0 +1,80 @@ +{ + "_name_or_path": "ntu-spml/distilhubert", + "activation_dropout": 0.1, + "apply_spec_augment": false, + "architectures": [ + "HubertForSequenceClassification" + ], + "attention_dropout": 0.1, + "bos_token_id": 1, + "classifier_proj_size": 256, + "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, + "do_stable_layer_norm": false, + "eos_token_id": 2, + "feat_extract_activation": "gelu", + "feat_extract_norm": "group", + "feat_proj_dropout": 0.0, + "feat_proj_layer_norm": false, + "final_dropout": 0.0, + "hidden_act": "gelu", + "hidden_dropout": 0.1, + "hidden_size": 768, + "id2label": { + "0": "crying", + "1": "no_crying" + }, + "initializer_range": 0.02, + "intermediate_size": 3072, + "label2id": { + "crying": "0", + "no_crying": "1" + }, + "layer_norm_eps": 1e-05, + "layerdrop": 0.0, + "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": 0.05, + "model_type": "hubert", + "num_attention_heads": 12, + "num_conv_pos_embedding_groups": 16, + "num_conv_pos_embeddings": 128, + "num_feat_extract_layers": 7, + "num_hidden_layers": 2, + "pad_token_id": 0, + "torch_dtype": "float32", + "transformers_version": "4.44.2", + "use_weighted_layer_sum": false, + "vocab_size": 32 +} diff --git a/checkpoint-114/model.safetensors b/checkpoint-114/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..f14770d0b87f2b26c6d201404ffa61659d28dd72 --- /dev/null +++ b/checkpoint-114/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b910bf6b8ec6e19d58a4b52c50de712e949fcdaeece42948112d2e5d95e8f47c +size 94763496 diff --git a/checkpoint-114/optimizer.pt b/checkpoint-114/optimizer.pt new file mode 100644 index 0000000000000000000000000000000000000000..5411201dd41a9064ffc7fac76b2946e89c057e63 --- /dev/null +++ b/checkpoint-114/optimizer.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:85b95917e0620b5556323a8c02ab048666c97ab19fb25f8fa5f1554c0bfc225f +size 189552570 diff --git a/checkpoint-114/preprocessor_config.json b/checkpoint-114/preprocessor_config.json new file mode 100644 index 0000000000000000000000000000000000000000..8ad80e2bc82023595cde1d13979733ba2dca60cc --- /dev/null +++ b/checkpoint-114/preprocessor_config.json @@ -0,0 +1,9 @@ +{ + "do_normalize": false, + "feature_extractor_type": "Wav2Vec2FeatureExtractor", + "feature_size": 1, + "padding_side": "right", + "padding_value": 0, + "return_attention_mask": false, + "sampling_rate": 16000 +} diff --git a/checkpoint-114/rng_state.pth b/checkpoint-114/rng_state.pth new file mode 100644 index 0000000000000000000000000000000000000000..21fe7fd38b05ffce4ceb1857615fd655a40117eb --- /dev/null +++ b/checkpoint-114/rng_state.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:636081736c072d4bb3052b5541d861668c6c103f9a3ad62f001239e9a3a36fb1 +size 14308 diff --git a/checkpoint-114/scheduler.pt b/checkpoint-114/scheduler.pt new file mode 100644 index 0000000000000000000000000000000000000000..b1e2009c54932eb6472d1f2fd62af51d5ece4cb2 --- /dev/null +++ b/checkpoint-114/scheduler.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3fde2346e00a3c4644832268d0eaddcd20d8548145016394246c018bb8a28ba1 +size 1064 diff --git a/checkpoint-114/trainer_state.json b/checkpoint-114/trainer_state.json new file mode 100644 index 0000000000000000000000000000000000000000..e6aff721b48bf663b175c0c174a59349b80bc665 --- /dev/null +++ b/checkpoint-114/trainer_state.json @@ -0,0 +1,150 @@ +{ + "best_metric": 0.9950980392156863, + "best_model_checkpoint": "distilhubert-finetuned-cry-detector/checkpoint-102", + "epoch": 8.941176470588236, + "eval_steps": 500, + "global_step": 114, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.9411764705882353, + "eval_accuracy": 0.9362745098039216, + "eval_f1": 0.9365356058160116, + "eval_loss": 0.19313116371631622, + "eval_precision": 0.9372234219385678, + "eval_recall": 0.9362745098039216, + "eval_runtime": 3.0986, + "eval_samples_per_second": 65.836, + "eval_steps_per_second": 8.391, + "step": 12 + }, + { + "epoch": 1.9607843137254903, + "eval_accuracy": 0.9705882352941176, + "eval_f1": 0.9704025549613785, + "eval_loss": 0.09495726972818375, + "eval_precision": 0.971039759187257, + "eval_recall": 0.9705882352941176, + "eval_runtime": 3.1171, + "eval_samples_per_second": 65.446, + "eval_steps_per_second": 8.341, + "step": 25 + }, + { + "epoch": 2.980392156862745, + "eval_accuracy": 0.9803921568627451, + "eval_f1": 0.9803921568627451, + "eval_loss": 0.06114682927727699, + "eval_precision": 0.9803921568627451, + "eval_recall": 0.9803921568627451, + "eval_runtime": 3.2163, + "eval_samples_per_second": 63.428, + "eval_steps_per_second": 8.084, + "step": 38 + }, + { + "epoch": 4.0, + "eval_accuracy": 0.9852941176470589, + "eval_f1": 0.9852722698274712, + "eval_loss": 0.049169763922691345, + "eval_precision": 0.9853018065952722, + "eval_recall": 0.9852941176470589, + "eval_runtime": 3.1223, + "eval_samples_per_second": 65.336, + "eval_steps_per_second": 8.327, + "step": 51 + }, + { + "epoch": 4.9411764705882355, + "eval_accuracy": 0.9803921568627451, + "eval_f1": 0.9804969685242518, + "eval_loss": 0.05879249796271324, + "eval_precision": 0.9813976872800402, + "eval_recall": 0.9803921568627451, + "eval_runtime": 3.1449, + "eval_samples_per_second": 64.868, + "eval_steps_per_second": 8.267, + "step": 63 + }, + { + "epoch": 5.96078431372549, + "eval_accuracy": 0.9852941176470589, + "eval_f1": 0.9853150765112866, + "eval_loss": 0.03680075705051422, + "eval_precision": 0.9853868369053048, + "eval_recall": 0.9852941176470589, + "eval_runtime": 3.1698, + "eval_samples_per_second": 64.358, + "eval_steps_per_second": 8.203, + "step": 76 + }, + { + "epoch": 6.980392156862745, + "eval_accuracy": 0.9901960784313726, + "eval_f1": 0.9901663478316141, + "eval_loss": 0.0382298119366169, + "eval_precision": 0.9903446226975638, + "eval_recall": 0.9901960784313726, + "eval_runtime": 3.1376, + "eval_samples_per_second": 65.019, + "eval_steps_per_second": 8.287, + "step": 89 + }, + { + "epoch": 8.0, + "eval_accuracy": 0.9950980392156863, + "eval_f1": 0.995090756609157, + "eval_loss": 0.03175761550664902, + "eval_precision": 0.995135458763658, + "eval_recall": 0.9950980392156863, + "eval_runtime": 2.949, + "eval_samples_per_second": 69.175, + "eval_steps_per_second": 8.816, + "step": 102 + }, + { + "epoch": 8.941176470588236, + "eval_accuracy": 0.9852941176470589, + "eval_f1": 0.9853150765112866, + "eval_loss": 0.033123087137937546, + "eval_precision": 0.9853868369053048, + "eval_recall": 0.9852941176470589, + "eval_runtime": 2.9458, + "eval_samples_per_second": 69.251, + "eval_steps_per_second": 8.826, + "step": 114 + } + ], + "logging_steps": 500, + "max_steps": 120, + "num_input_tokens_seen": 0, + "num_train_epochs": 10, + "save_steps": 500, + "stateful_callbacks": { + "EarlyStoppingCallback": { + "args": { + "early_stopping_patience": 3, + "early_stopping_threshold": 0.0 + }, + "attributes": { + "early_stopping_patience_counter": 0 + } + }, + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": false + }, + "attributes": {} + } + }, + "total_flos": 1.1834894644418458e+17, + "train_batch_size": 8, + "trial_name": null, + "trial_params": null +} diff --git a/checkpoint-114/training_args.bin b/checkpoint-114/training_args.bin new file mode 100644 index 0000000000000000000000000000000000000000..ff5c682679da830384645399a104cd532805ce20 --- /dev/null +++ b/checkpoint-114/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3b77b2a33ca88330b7909aacab81332695e9c9dcf797e7a7757ca31ffcde6f31 +size 5240 diff --git a/checkpoint-12/model.safetensors b/checkpoint-12/model.safetensors index 99bb8b757cc82dc894263c1ed673d5ba4340cbe6..40d16723570a92e46896da19a65dff0193a0db0b 100644 --- a/checkpoint-12/model.safetensors +++ b/checkpoint-12/model.safetensors @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:ed24ddc80fbfa0595409861cddf08e518561fbdd7c4461e5edac2ef1da033508 +oid sha256:7ac774d14fe5167fbe1adddc20941f151440dc32f43d6762bbe9d60a26726e8e size 94763496 diff --git a/checkpoint-12/optimizer.pt b/checkpoint-12/optimizer.pt index 985b2bf97b1cc9344a23711f2501b203e30ac596..811d8ee158681897cbda6dcfeb3fced7ceb9ac59 100644 --- a/checkpoint-12/optimizer.pt +++ b/checkpoint-12/optimizer.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:cf1fd2bcc1d5386abd738663f70a4cfa450e0389d1bb46d5ebbd8d22ee0082c5 +oid sha256:56315b610ef113026417660dc64000a73d31bcdf367d4ade81935feefcbbb425 size 189552570 diff --git a/checkpoint-12/rng_state.pth b/checkpoint-12/rng_state.pth index f4a82f4148b7f32e7ca96d25f7ebd274acbd3675..66f6550a5597b89489b605bce6f9fb3d877d7aba 100644 --- a/checkpoint-12/rng_state.pth +++ b/checkpoint-12/rng_state.pth @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:0dd5e5067cb0b69c310c1cb6eb41548b436585db6c11789bed910bb1926cd7c4 +oid sha256:caebb0abc1e3f29ed8c6d6714ac9331ab5a6b59efbd9b310598bb00f5bca4c19 size 14308 diff --git a/checkpoint-12/scheduler.pt b/checkpoint-12/scheduler.pt index 036937fe8481b4b940b667cedfea07446d4a8b9f..58161fb1369f7d88952f1089463d3b56c2d170b0 100644 --- a/checkpoint-12/scheduler.pt +++ b/checkpoint-12/scheduler.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:00ac470d61d5b4613444f686474fe5bf71dbe645a07555e3c0011199a78b17ab +oid sha256:4ca0321d043e24273cc9815784db5e90a317cada9c48cd156d448ab1591d936c size 1064 diff --git a/checkpoint-12/trainer_state.json b/checkpoint-12/trainer_state.json index aebb202a00a7fc468c005490cf503831761628fd..785209b7e3af6b6d98a9fd6f7e7ca36da6711640 100644 --- a/checkpoint-12/trainer_state.json +++ b/checkpoint-12/trainer_state.json @@ -1,5 +1,5 @@ { - "best_metric": 0.9558823529411765, + "best_metric": 0.9362745098039216, "best_model_checkpoint": "distilhubert-finetuned-cry-detector/checkpoint-12", "epoch": 0.9411764705882353, "eval_steps": 500, @@ -10,21 +10,21 @@ "log_history": [ { "epoch": 0.9411764705882353, - "eval_accuracy": 0.9558823529411765, - "eval_f1": 0.9555272696227192, - "eval_loss": 0.33786964416503906, - "eval_precision": 0.9564428025638596, - "eval_recall": 0.9558823529411765, - "eval_runtime": 3.0231, - "eval_samples_per_second": 67.479, - "eval_steps_per_second": 8.6, + "eval_accuracy": 0.9362745098039216, + "eval_f1": 0.9365356058160116, + "eval_loss": 0.19313116371631622, + "eval_precision": 0.9372234219385678, + "eval_recall": 0.9362745098039216, + "eval_runtime": 3.0986, + "eval_samples_per_second": 65.836, + "eval_steps_per_second": 8.391, "step": 12 } ], "logging_steps": 500, - "max_steps": 12, + "max_steps": 120, "num_input_tokens_seen": 0, - "num_train_epochs": 1, + "num_train_epochs": 10, "save_steps": 500, "stateful_callbacks": { "EarlyStoppingCallback": { @@ -42,12 +42,12 @@ "should_evaluate": false, "should_log": false, "should_save": true, - "should_training_stop": true + "should_training_stop": false }, "attributes": {} } }, - "total_flos": 1.238025689554944e+16, + "total_flos": 1.315729776671232e+16, "train_batch_size": 8, "trial_name": null, "trial_params": null diff --git a/checkpoint-12/training_args.bin b/checkpoint-12/training_args.bin index 17bea61543f565c72b3b723abec4a0277f862e92..ff5c682679da830384645399a104cd532805ce20 100644 --- a/checkpoint-12/training_args.bin +++ b/checkpoint-12/training_args.bin @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:a1c27b4ecab2c1ddc610f4ce2ead133759bcabe788a27ce90b2b23560c4e6ad3 +oid sha256:3b77b2a33ca88330b7909aacab81332695e9c9dcf797e7a7757ca31ffcde6f31 size 5240 diff --git a/checkpoint-120/config.json b/checkpoint-120/config.json new file mode 100644 index 0000000000000000000000000000000000000000..4550a22abfa803293b5dc6b2633d0201ac2ff44d --- /dev/null +++ b/checkpoint-120/config.json @@ -0,0 +1,80 @@ +{ + "_name_or_path": "ntu-spml/distilhubert", + "activation_dropout": 0.1, + "apply_spec_augment": false, + "architectures": [ + "HubertForSequenceClassification" + ], + "attention_dropout": 0.1, + "bos_token_id": 1, + "classifier_proj_size": 256, + "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, + "do_stable_layer_norm": false, + "eos_token_id": 2, + "feat_extract_activation": "gelu", + "feat_extract_norm": "group", + "feat_proj_dropout": 0.0, + "feat_proj_layer_norm": false, + "final_dropout": 0.0, + "hidden_act": "gelu", + "hidden_dropout": 0.1, + "hidden_size": 768, + "id2label": { + "0": "crying", + "1": "no_crying" + }, + "initializer_range": 0.02, + "intermediate_size": 3072, + "label2id": { + "crying": "0", + "no_crying": "1" + }, + "layer_norm_eps": 1e-05, + "layerdrop": 0.0, + "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": 0.05, + "model_type": "hubert", + "num_attention_heads": 12, + "num_conv_pos_embedding_groups": 16, + "num_conv_pos_embeddings": 128, + "num_feat_extract_layers": 7, + "num_hidden_layers": 2, + "pad_token_id": 0, + "torch_dtype": "float32", + "transformers_version": "4.44.2", + "use_weighted_layer_sum": false, + "vocab_size": 32 +} diff --git a/checkpoint-120/model.safetensors b/checkpoint-120/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..9013a72a6276cab9ffb6eb3390d4d615e1dc081b --- /dev/null +++ b/checkpoint-120/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3258e17c3151e1d9899b5ecaa38ee7208137f2c975f17a7f66834494fb6f1ddb +size 94763496 diff --git a/checkpoint-120/optimizer.pt b/checkpoint-120/optimizer.pt new file mode 100644 index 0000000000000000000000000000000000000000..92ddef1f87f8067591bb67a9948bec3555f5ad28 --- /dev/null +++ b/checkpoint-120/optimizer.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d67e947c82c7eff62fe82e2161b429d0f674dcb710ae9e066c9aaeaa53719096 +size 189552570 diff --git a/checkpoint-120/preprocessor_config.json b/checkpoint-120/preprocessor_config.json new file mode 100644 index 0000000000000000000000000000000000000000..8ad80e2bc82023595cde1d13979733ba2dca60cc --- /dev/null +++ b/checkpoint-120/preprocessor_config.json @@ -0,0 +1,9 @@ +{ + "do_normalize": false, + "feature_extractor_type": "Wav2Vec2FeatureExtractor", + "feature_size": 1, + "padding_side": "right", + "padding_value": 0, + "return_attention_mask": false, + "sampling_rate": 16000 +} diff --git a/checkpoint-120/rng_state.pth b/checkpoint-120/rng_state.pth new file mode 100644 index 0000000000000000000000000000000000000000..d109978affbdd803c224b72ce809580eaf597efa --- /dev/null +++ b/checkpoint-120/rng_state.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:286671099adfaa4371e8fead24d8a5f622004012b7f075f68979922946bd3511 +size 14308 diff --git a/checkpoint-120/scheduler.pt b/checkpoint-120/scheduler.pt new file mode 100644 index 0000000000000000000000000000000000000000..e376437090a954f14b92546ec0d65c1e1d0f7fcd --- /dev/null +++ b/checkpoint-120/scheduler.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e2409fb1c3fd1c2de2c63b5459c49a97abc9a50084af4337cc5090c00ef975a6 +size 1064 diff --git a/checkpoint-120/trainer_state.json b/checkpoint-120/trainer_state.json new file mode 100644 index 0000000000000000000000000000000000000000..cd83a4f967a1265f5cfce63143ea8a90229f3ed9 --- /dev/null +++ b/checkpoint-120/trainer_state.json @@ -0,0 +1,162 @@ +{ + "best_metric": 0.9950980392156863, + "best_model_checkpoint": "distilhubert-finetuned-cry-detector/checkpoint-102", + "epoch": 9.411764705882353, + "eval_steps": 500, + "global_step": 120, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.9411764705882353, + "eval_accuracy": 0.9362745098039216, + "eval_f1": 0.9365356058160116, + "eval_loss": 0.19313116371631622, + "eval_precision": 0.9372234219385678, + "eval_recall": 0.9362745098039216, + "eval_runtime": 3.0986, + "eval_samples_per_second": 65.836, + "eval_steps_per_second": 8.391, + "step": 12 + }, + { + "epoch": 1.9607843137254903, + "eval_accuracy": 0.9705882352941176, + "eval_f1": 0.9704025549613785, + "eval_loss": 0.09495726972818375, + "eval_precision": 0.971039759187257, + "eval_recall": 0.9705882352941176, + "eval_runtime": 3.1171, + "eval_samples_per_second": 65.446, + "eval_steps_per_second": 8.341, + "step": 25 + }, + { + "epoch": 2.980392156862745, + "eval_accuracy": 0.9803921568627451, + "eval_f1": 0.9803921568627451, + "eval_loss": 0.06114682927727699, + "eval_precision": 0.9803921568627451, + "eval_recall": 0.9803921568627451, + "eval_runtime": 3.2163, + "eval_samples_per_second": 63.428, + "eval_steps_per_second": 8.084, + "step": 38 + }, + { + "epoch": 4.0, + "eval_accuracy": 0.9852941176470589, + "eval_f1": 0.9852722698274712, + "eval_loss": 0.049169763922691345, + "eval_precision": 0.9853018065952722, + "eval_recall": 0.9852941176470589, + "eval_runtime": 3.1223, + "eval_samples_per_second": 65.336, + "eval_steps_per_second": 8.327, + "step": 51 + }, + { + "epoch": 4.9411764705882355, + "eval_accuracy": 0.9803921568627451, + "eval_f1": 0.9804969685242518, + "eval_loss": 0.05879249796271324, + "eval_precision": 0.9813976872800402, + "eval_recall": 0.9803921568627451, + "eval_runtime": 3.1449, + "eval_samples_per_second": 64.868, + "eval_steps_per_second": 8.267, + "step": 63 + }, + { + "epoch": 5.96078431372549, + "eval_accuracy": 0.9852941176470589, + "eval_f1": 0.9853150765112866, + "eval_loss": 0.03680075705051422, + "eval_precision": 0.9853868369053048, + "eval_recall": 0.9852941176470589, + "eval_runtime": 3.1698, + "eval_samples_per_second": 64.358, + "eval_steps_per_second": 8.203, + "step": 76 + }, + { + "epoch": 6.980392156862745, + "eval_accuracy": 0.9901960784313726, + "eval_f1": 0.9901663478316141, + "eval_loss": 0.0382298119366169, + "eval_precision": 0.9903446226975638, + "eval_recall": 0.9901960784313726, + "eval_runtime": 3.1376, + "eval_samples_per_second": 65.019, + "eval_steps_per_second": 8.287, + "step": 89 + }, + { + "epoch": 8.0, + "eval_accuracy": 0.9950980392156863, + "eval_f1": 0.995090756609157, + "eval_loss": 0.03175761550664902, + "eval_precision": 0.995135458763658, + "eval_recall": 0.9950980392156863, + "eval_runtime": 2.949, + "eval_samples_per_second": 69.175, + "eval_steps_per_second": 8.816, + "step": 102 + }, + { + "epoch": 8.941176470588236, + "eval_accuracy": 0.9852941176470589, + "eval_f1": 0.9853150765112866, + "eval_loss": 0.033123087137937546, + "eval_precision": 0.9853868369053048, + "eval_recall": 0.9852941176470589, + "eval_runtime": 2.9458, + "eval_samples_per_second": 69.251, + "eval_steps_per_second": 8.826, + "step": 114 + }, + { + "epoch": 9.411764705882353, + "eval_accuracy": 0.9852941176470589, + "eval_f1": 0.9853150765112866, + "eval_loss": 0.033173661679029465, + "eval_precision": 0.9853868369053048, + "eval_recall": 0.9852941176470589, + "eval_runtime": 2.9666, + "eval_samples_per_second": 68.766, + "eval_steps_per_second": 8.764, + "step": 120 + } + ], + "logging_steps": 500, + "max_steps": 120, + "num_input_tokens_seen": 0, + "num_train_epochs": 10, + "save_steps": 500, + "stateful_callbacks": { + "EarlyStoppingCallback": { + "args": { + "early_stopping_patience": 3, + "early_stopping_threshold": 0.0 + }, + "attributes": { + "early_stopping_patience_counter": 0 + } + }, + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": true + }, + "attributes": {} + } + }, + "total_flos": 1.237670053721906e+17, + "train_batch_size": 8, + "trial_name": null, + "trial_params": null +} diff --git a/checkpoint-120/training_args.bin b/checkpoint-120/training_args.bin new file mode 100644 index 0000000000000000000000000000000000000000..ff5c682679da830384645399a104cd532805ce20 --- /dev/null +++ b/checkpoint-120/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3b77b2a33ca88330b7909aacab81332695e9c9dcf797e7a7757ca31ffcde6f31 +size 5240 diff --git a/checkpoint-25/model.safetensors b/checkpoint-25/model.safetensors index ad2ed673dfba02cb6352ec93279cd75b4822fdda..c4105d3aa1b23dbb741129bad9f1427b305f01e0 100644 --- a/checkpoint-25/model.safetensors +++ b/checkpoint-25/model.safetensors @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:457aaf1ca73f2bd8fdf652ff897b1db85473a23401cdea192ccda04c31e902f9 +oid sha256:da1949ab9da06d9f94ddb053df98c408f27159d75669cf1d2eb4e1cc47323344 size 94763496 diff --git a/checkpoint-25/optimizer.pt b/checkpoint-25/optimizer.pt index 714998c128a7af170883ab6bd68916f961bcabe8..bca5d3bb40036114acfbbbbef1c66e25db113c41 100644 --- a/checkpoint-25/optimizer.pt +++ b/checkpoint-25/optimizer.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:0f78eeb88abcee76fe1f3cb427847a3366d60d465ad4933bdf631baef11c677d +oid sha256:64fe5dbcc580831e63b26cf1b5d7e9409f5b456d1141636612d7fca115a01a87 size 189552570 diff --git a/checkpoint-25/rng_state.pth b/checkpoint-25/rng_state.pth index 52bba65b4817a8be78c3aa5195a21ce7895d0a29..1c4c63dda4a854e1c26897664f79ed87026fec03 100644 --- a/checkpoint-25/rng_state.pth +++ b/checkpoint-25/rng_state.pth @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:8e5c954d42fa0d2b945cf8a644f5bb95c0ecae8f78ef6f89815808920b5b330c -size 14244 +oid sha256:272c079a133d684090b7a24b10497f4bd703dbf9ecfecafe1a788c62b4404ce6 +size 14308 diff --git a/checkpoint-25/scheduler.pt b/checkpoint-25/scheduler.pt index f3d5446c04ea64775d05571d500507a776186b13..9ef00324d1a89eb242062734596c5a6bbccbe979 100644 --- a/checkpoint-25/scheduler.pt +++ b/checkpoint-25/scheduler.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:91532783b102a4c8ff5ee86df37d3662ca7a62e198729a349c132e2c9a68d5d0 +oid sha256:6633f18d367676ea03844228fc7112a6a9464032824bb5f9f17a85ee888f12e8 size 1064 diff --git a/checkpoint-25/trainer_state.json b/checkpoint-25/trainer_state.json index e5482635e86237a6a41cb25830d9debd66900385..273a13e76390f6d439d7b36fe30982be18f30f33 100644 --- a/checkpoint-25/trainer_state.json +++ b/checkpoint-25/trainer_state.json @@ -1,5 +1,5 @@ { - "best_metric": 0.9558823529411765, + "best_metric": 0.9705882352941176, "best_model_checkpoint": "distilhubert-finetuned-cry-detector/checkpoint-25", "epoch": 1.9607843137254903, "eval_steps": 500, @@ -10,33 +10,33 @@ "log_history": [ { "epoch": 0.9411764705882353, - "eval_accuracy": 0.9509803921568627, - "eval_f1": 0.9509803921568627, - "eval_loss": 0.19337645173072815, - "eval_precision": 0.9509803921568627, - "eval_recall": 0.9509803921568627, - "eval_runtime": 2.9937, - "eval_samples_per_second": 68.144, - "eval_steps_per_second": 8.685, + "eval_accuracy": 0.9362745098039216, + "eval_f1": 0.9365356058160116, + "eval_loss": 0.19313116371631622, + "eval_precision": 0.9372234219385678, + "eval_recall": 0.9362745098039216, + "eval_runtime": 3.0986, + "eval_samples_per_second": 65.836, + "eval_steps_per_second": 8.391, "step": 12 }, { "epoch": 1.9607843137254903, - "eval_accuracy": 0.9558823529411765, - "eval_f1": 0.9558168094824137, - "eval_loss": 0.11207430809736252, - "eval_precision": 0.9558008500901145, - "eval_recall": 0.9558823529411765, - "eval_runtime": 2.9789, - "eval_samples_per_second": 68.481, - "eval_steps_per_second": 8.728, + "eval_accuracy": 0.9705882352941176, + "eval_f1": 0.9704025549613785, + "eval_loss": 0.09495726972818375, + "eval_precision": 0.971039759187257, + "eval_recall": 0.9705882352941176, + "eval_runtime": 3.1171, + "eval_samples_per_second": 65.446, + "eval_steps_per_second": 8.341, "step": 25 } ], "logging_steps": 500, - "max_steps": 48, + "max_steps": 120, "num_input_tokens_seen": 0, - "num_train_epochs": 4, + "num_train_epochs": 10, "save_steps": 500, "stateful_callbacks": { "EarlyStoppingCallback": { @@ -59,7 +59,7 @@ "attributes": {} } }, - "total_flos": 2.632285534507315e+16, + "total_flos": 2.630848254506803e+16, "train_batch_size": 8, "trial_name": null, "trial_params": null diff --git a/checkpoint-25/training_args.bin b/checkpoint-25/training_args.bin index 80cc3d9fd69a2c32d45e3859728c874d2a0d2bd5..ff5c682679da830384645399a104cd532805ce20 100644 --- a/checkpoint-25/training_args.bin +++ b/checkpoint-25/training_args.bin @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:b08c1ef2199623b781e926f1d827964186fca47326620731302bfe37c5d058b9 +oid sha256:3b77b2a33ca88330b7909aacab81332695e9c9dcf797e7a7757ca31ffcde6f31 size 5240 diff --git a/checkpoint-38/model.safetensors b/checkpoint-38/model.safetensors index 0dddf52a56a966c75d0b19e1a441bd0a5440add5..68a61bbbeb38439c8763713dc9eaceb162366af7 100644 --- a/checkpoint-38/model.safetensors +++ b/checkpoint-38/model.safetensors @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:365a9fb794f55008ff39b21086b11700c109a194efb2127a6db2145604a614b6 +oid sha256:d8bc28d8ebd1dbdd73656a784624c72928eb10e21d9e6acdce2e1d312d787010 size 94763496 diff --git a/checkpoint-38/optimizer.pt b/checkpoint-38/optimizer.pt index ed6c18f8824fc13c7f975a0dd2d8c9c163b2bf9c..978d478f0ac60ba62d9c9a4a416faf1da54a3b0c 100644 --- a/checkpoint-38/optimizer.pt +++ b/checkpoint-38/optimizer.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:198372300c7cb71ea5cd197e4b8f17970da74cfc971f4a284751e8c5c2fe5397 +oid sha256:0ee1b6c4df5d7a5ef50f7350cb964bab9875222b057b8eb71768826fceb9156f size 189552570 diff --git a/checkpoint-38/rng_state.pth b/checkpoint-38/rng_state.pth index 8122fb10ab702c889a2f1e5e133327682c38e92a..07d6f84a9330a8beddd35d4bf156f27a5ffed9b6 100644 --- a/checkpoint-38/rng_state.pth +++ b/checkpoint-38/rng_state.pth @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:780c51c50482e6e8f628b4687957abc3bb4ff7f96b977c02d8538cf6f0d2d494 -size 14244 +oid sha256:4d19f57e20f91b35ae6289880b02cdfa14033ed6b20b4d1730bfb179a9264102 +size 14308 diff --git a/checkpoint-38/scheduler.pt b/checkpoint-38/scheduler.pt index a4bf89e18cca0e1fe8963f3c9d52cbda907cb3c0..ddd422c54efec4259253d776031e258b3b987bca 100644 --- a/checkpoint-38/scheduler.pt +++ b/checkpoint-38/scheduler.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:96ddec5715a7e2a3cc7175b7f159a7e1fd2303094015b4bffdae6a236e889974 +oid sha256:5df84074f7631c8ca35330b9e20d4190996409e954261c1b476ad0fabe74b93b size 1064 diff --git a/checkpoint-38/trainer_state.json b/checkpoint-38/trainer_state.json index bc1131eb374ceb96701474e75122ee07323f8d83..bd6b3186b43826198c82b2002893ac03ad6dc677 100644 --- a/checkpoint-38/trainer_state.json +++ b/checkpoint-38/trainer_state.json @@ -1,5 +1,5 @@ { - "best_metric": 0.9607843137254902, + "best_metric": 0.9803921568627451, "best_model_checkpoint": "distilhubert-finetuned-cry-detector/checkpoint-38", "epoch": 2.980392156862745, "eval_steps": 500, @@ -10,45 +10,45 @@ "log_history": [ { "epoch": 0.9411764705882353, - "eval_accuracy": 0.9509803921568627, - "eval_f1": 0.9509803921568627, - "eval_loss": 0.19337645173072815, - "eval_precision": 0.9509803921568627, - "eval_recall": 0.9509803921568627, - "eval_runtime": 2.9937, - "eval_samples_per_second": 68.144, - "eval_steps_per_second": 8.685, + "eval_accuracy": 0.9362745098039216, + "eval_f1": 0.9365356058160116, + "eval_loss": 0.19313116371631622, + "eval_precision": 0.9372234219385678, + "eval_recall": 0.9362745098039216, + "eval_runtime": 3.0986, + "eval_samples_per_second": 65.836, + "eval_steps_per_second": 8.391, "step": 12 }, { "epoch": 1.9607843137254903, - "eval_accuracy": 0.9558823529411765, - "eval_f1": 0.9558168094824137, - "eval_loss": 0.11207430809736252, - "eval_precision": 0.9558008500901145, - "eval_recall": 0.9558823529411765, - "eval_runtime": 2.9789, - "eval_samples_per_second": 68.481, - "eval_steps_per_second": 8.728, + "eval_accuracy": 0.9705882352941176, + "eval_f1": 0.9704025549613785, + "eval_loss": 0.09495726972818375, + "eval_precision": 0.971039759187257, + "eval_recall": 0.9705882352941176, + "eval_runtime": 3.1171, + "eval_samples_per_second": 65.446, + "eval_steps_per_second": 8.341, "step": 25 }, { "epoch": 2.980392156862745, - "eval_accuracy": 0.9607843137254902, - "eval_f1": 0.9608937528499771, - "eval_loss": 0.08983828127384186, - "eval_precision": 0.9611995291537668, - "eval_recall": 0.9607843137254902, - "eval_runtime": 2.927, - "eval_samples_per_second": 69.695, - "eval_steps_per_second": 8.883, + "eval_accuracy": 0.9803921568627451, + "eval_f1": 0.9803921568627451, + "eval_loss": 0.06114682927727699, + "eval_precision": 0.9803921568627451, + "eval_recall": 0.9803921568627451, + "eval_runtime": 3.2163, + "eval_samples_per_second": 63.428, + "eval_steps_per_second": 8.084, "step": 38 } ], "logging_steps": 500, - "max_steps": 48, + "max_steps": 120, "num_input_tokens_seen": 0, - "num_train_epochs": 4, + "num_train_epochs": 10, "save_steps": 500, "stateful_callbacks": { "EarlyStoppingCallback": { @@ -71,7 +71,7 @@ "attributes": {} } }, - "total_flos": 3.947040143988326e+16, + "total_flos": 3.946056334924685e+16, "train_batch_size": 8, "trial_name": null, "trial_params": null diff --git a/checkpoint-38/training_args.bin b/checkpoint-38/training_args.bin index 80cc3d9fd69a2c32d45e3859728c874d2a0d2bd5..ff5c682679da830384645399a104cd532805ce20 100644 --- a/checkpoint-38/training_args.bin +++ b/checkpoint-38/training_args.bin @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:b08c1ef2199623b781e926f1d827964186fca47326620731302bfe37c5d058b9 +oid sha256:3b77b2a33ca88330b7909aacab81332695e9c9dcf797e7a7757ca31ffcde6f31 size 5240 diff --git a/checkpoint-51/model.safetensors b/checkpoint-51/model.safetensors index 2e227644384b23941e53332db814bf965f9ef5a4..112858a8afa8188ad6050f50ea95b9944ed3004c 100644 --- a/checkpoint-51/model.safetensors +++ b/checkpoint-51/model.safetensors @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:8e9b6729c50f84d9b261120a8e72c4b3a3da69e27bb43431fdcd37c673c45006 +oid sha256:072781416d7434132c8e1b73276d4503dca0b1b8db503afa1a8ba07bc69b3ae9 size 94763496 diff --git a/checkpoint-51/optimizer.pt b/checkpoint-51/optimizer.pt index d07320368ba3c68eafba5762e999969401dff021..1136b3e9ce2d8eaf2749dcfaade35f71f3f1dd90 100644 --- a/checkpoint-51/optimizer.pt +++ b/checkpoint-51/optimizer.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:461ec76d1bc7a1e71c52a5302f7d5be7e5eec1ad9fc45cdec1c8d7a0e5e545c1 +oid sha256:da71c2ca09d32a4c673ba10e3259d6198c7a2fbaa9004a4a0eaa7d9cc8d8fecb size 189552570 diff --git a/checkpoint-51/rng_state.pth b/checkpoint-51/rng_state.pth index 92119e01ba23370adf79d5a4db474dd5ed91439a..6c5779780a29ba2b20a639fe2bee221e351a122b 100644 --- a/checkpoint-51/rng_state.pth +++ b/checkpoint-51/rng_state.pth @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:8bfb428aabf8fc7d0150525679311cfec661fbbda39191a37add4ec1737f1b54 -size 14244 +oid sha256:bdc415b22b0f41c29b31b3bf40c83a0c3865a35793ba92c48d3d575e69ed5279 +size 14308 diff --git a/checkpoint-51/scheduler.pt b/checkpoint-51/scheduler.pt index d8e27ccd4b43209e1caa572dbd3c40faa48bdf3d..6304625c226daa5a48682874226d5b29f78d2934 100644 --- a/checkpoint-51/scheduler.pt +++ b/checkpoint-51/scheduler.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:bec8121eea0c42c283a1e74848057f68b938b9185ee7ce8d2f0f77b15cf4fe84 +oid sha256:99f2e94f9942cdbd4528ad8ed9baf4498c04d7503c2ef6dab9521dfcb4021a49 size 1064 diff --git a/checkpoint-51/trainer_state.json b/checkpoint-51/trainer_state.json index 7a642ac932785c9c44916dffa441b8cf68bfc2cc..e46cbe932239ca0fad89a43ced5f8a8c84bfacd5 100644 --- a/checkpoint-51/trainer_state.json +++ b/checkpoint-51/trainer_state.json @@ -1,5 +1,5 @@ { - "best_metric": 0.9705882352941176, + "best_metric": 0.9852941176470589, "best_model_checkpoint": "distilhubert-finetuned-cry-detector/checkpoint-51", "epoch": 4.0, "eval_steps": 500, @@ -10,57 +10,57 @@ "log_history": [ { "epoch": 0.9411764705882353, - "eval_accuracy": 0.9558823529411765, - "eval_f1": 0.9558168094824137, - "eval_loss": 0.18768352270126343, - "eval_precision": 0.9558008500901145, - "eval_recall": 0.9558823529411765, - "eval_runtime": 3.1767, - "eval_samples_per_second": 64.218, - "eval_steps_per_second": 8.185, + "eval_accuracy": 0.9362745098039216, + "eval_f1": 0.9365356058160116, + "eval_loss": 0.19313116371631622, + "eval_precision": 0.9372234219385678, + "eval_recall": 0.9362745098039216, + "eval_runtime": 3.0986, + "eval_samples_per_second": 65.836, + "eval_steps_per_second": 8.391, "step": 12 }, { "epoch": 1.9607843137254903, - "eval_accuracy": 0.9607843137254902, - "eval_f1": 0.9603980954872683, - "eval_loss": 0.10839918255805969, - "eval_precision": 0.961865628604383, - "eval_recall": 0.9607843137254902, - "eval_runtime": 3.1396, - "eval_samples_per_second": 64.977, - "eval_steps_per_second": 8.281, + "eval_accuracy": 0.9705882352941176, + "eval_f1": 0.9704025549613785, + "eval_loss": 0.09495726972818375, + "eval_precision": 0.971039759187257, + "eval_recall": 0.9705882352941176, + "eval_runtime": 3.1171, + "eval_samples_per_second": 65.446, + "eval_steps_per_second": 8.341, "step": 25 }, { "epoch": 2.980392156862745, - "eval_accuracy": 0.9607843137254902, - "eval_f1": 0.9608937528499771, - "eval_loss": 0.07022649049758911, - "eval_precision": 0.9611995291537668, - "eval_recall": 0.9607843137254902, - "eval_runtime": 3.1392, - "eval_samples_per_second": 64.985, - "eval_steps_per_second": 8.282, + "eval_accuracy": 0.9803921568627451, + "eval_f1": 0.9803921568627451, + "eval_loss": 0.06114682927727699, + "eval_precision": 0.9803921568627451, + "eval_recall": 0.9803921568627451, + "eval_runtime": 3.2163, + "eval_samples_per_second": 63.428, + "eval_steps_per_second": 8.084, "step": 38 }, { "epoch": 4.0, - "eval_accuracy": 0.9705882352941176, - "eval_f1": 0.9707454527863777, - "eval_loss": 0.05151732638478279, - "eval_precision": 0.9716895305130601, - "eval_recall": 0.9705882352941176, - "eval_runtime": 3.1558, - "eval_samples_per_second": 64.644, - "eval_steps_per_second": 8.239, + "eval_accuracy": 0.9852941176470589, + "eval_f1": 0.9852722698274712, + "eval_loss": 0.049169763922691345, + "eval_precision": 0.9853018065952722, + "eval_recall": 0.9852941176470589, + "eval_runtime": 3.1223, + "eval_samples_per_second": 65.336, + "eval_steps_per_second": 8.327, "step": 51 } ], "logging_steps": 500, - "max_steps": 96, + "max_steps": 120, "num_input_tokens_seen": 0, - "num_train_epochs": 8, + "num_train_epochs": 10, "save_steps": 500, "stateful_callbacks": { "EarlyStoppingCallback": { @@ -83,7 +83,7 @@ "attributes": {} } }, - "total_flos": 5.25877328162016e+16, + "total_flos": 5.261731531342733e+16, "train_batch_size": 8, "trial_name": null, "trial_params": null diff --git a/checkpoint-51/training_args.bin b/checkpoint-51/training_args.bin index c9ee3705505943fb60df784486d5cefbd8aa885b..ff5c682679da830384645399a104cd532805ce20 100644 --- a/checkpoint-51/training_args.bin +++ b/checkpoint-51/training_args.bin @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:ecc689edbe3573cb27ca0a4ddd21670a9f672918abacfb53cac5b55b680695a7 +oid sha256:3b77b2a33ca88330b7909aacab81332695e9c9dcf797e7a7757ca31ffcde6f31 size 5240 diff --git a/checkpoint-63/model.safetensors b/checkpoint-63/model.safetensors index 89cf01c0dedf2321d40b75264d3df4cb8c97aa80..0c049aa02c2846b4f9e432a07a7fc79e6fa56e01 100644 --- a/checkpoint-63/model.safetensors +++ b/checkpoint-63/model.safetensors @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:fc362a1585df3481e2dcd870193d28d930ae325d6150daaaad1493b1996de42f +oid sha256:2c1568086524c8dced6705098aa1db3285da46fd4a5391fa6b9d6eb7df4502d4 size 94763496 diff --git a/checkpoint-63/optimizer.pt b/checkpoint-63/optimizer.pt index 5029e7942ea005bd6348b01a9dc457ed56fe63f0..8ee9d1eb536e5186df48146283a162dbc7aceb4d 100644 --- a/checkpoint-63/optimizer.pt +++ b/checkpoint-63/optimizer.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:61e3027407b1392589f3f7539771e589bd5fcb4b9cdc5fd8b3490f1845321f8d +oid sha256:b2e555fd925a25201dcce20a9e5cd0b805cd4f7aab958c817c9cb766841df306 size 189552570 diff --git a/checkpoint-63/rng_state.pth b/checkpoint-63/rng_state.pth index 280091fca239de26c5f86c04eb28e8df7c9cd3c7..2773df89225cc9185768aa4bebf56f78db3158ac 100644 --- a/checkpoint-63/rng_state.pth +++ b/checkpoint-63/rng_state.pth @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:fdfe969ed1ae8b4a9c1ad25a29ccab3cdc0a5691a8f2bbd3502eb40a686479b1 -size 14244 +oid sha256:fa1840af3a88fb6eeb907f8cfb790ef5793c08626b781f2e2ddc72a220df164a +size 14308 diff --git a/checkpoint-63/scheduler.pt b/checkpoint-63/scheduler.pt index 74c77eba587ec7fb2694dc3116d43fd5a1dc5012..24739529339e5791ed2aa200d7dfd611d30ef67d 100644 --- a/checkpoint-63/scheduler.pt +++ b/checkpoint-63/scheduler.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:af5dcadeffd3263bbd3829423086e16b3446f3994acc9aaa76e3c47c3489ed6a +oid sha256:d312d5601acda40460560c83523859ed8766ca9e53405a6efffea4f7db6bdf34 size 1064 diff --git a/checkpoint-63/trainer_state.json b/checkpoint-63/trainer_state.json index 485a8f3a2481ced487cde44d89486b61d2bd84f4..d54ac55f2d254a3b4e10361740324a5cf258298c 100644 --- a/checkpoint-63/trainer_state.json +++ b/checkpoint-63/trainer_state.json @@ -1,6 +1,6 @@ { - "best_metric": 0.9754901960784313, - "best_model_checkpoint": "distilhubert-finetuned-cry-detector/checkpoint-63", + "best_metric": 0.9852941176470589, + "best_model_checkpoint": "distilhubert-finetuned-cry-detector/checkpoint-51", "epoch": 4.9411764705882355, "eval_steps": 500, "global_step": 63, @@ -10,69 +10,69 @@ "log_history": [ { "epoch": 0.9411764705882353, - "eval_accuracy": 0.9558823529411765, - "eval_f1": 0.9558168094824137, - "eval_loss": 0.18768352270126343, - "eval_precision": 0.9558008500901145, - "eval_recall": 0.9558823529411765, - "eval_runtime": 3.1767, - "eval_samples_per_second": 64.218, - "eval_steps_per_second": 8.185, + "eval_accuracy": 0.9362745098039216, + "eval_f1": 0.9365356058160116, + "eval_loss": 0.19313116371631622, + "eval_precision": 0.9372234219385678, + "eval_recall": 0.9362745098039216, + "eval_runtime": 3.0986, + "eval_samples_per_second": 65.836, + "eval_steps_per_second": 8.391, "step": 12 }, { "epoch": 1.9607843137254903, - "eval_accuracy": 0.9607843137254902, - "eval_f1": 0.9603980954872683, - "eval_loss": 0.10839918255805969, - "eval_precision": 0.961865628604383, - "eval_recall": 0.9607843137254902, - "eval_runtime": 3.1396, - "eval_samples_per_second": 64.977, - "eval_steps_per_second": 8.281, + "eval_accuracy": 0.9705882352941176, + "eval_f1": 0.9704025549613785, + "eval_loss": 0.09495726972818375, + "eval_precision": 0.971039759187257, + "eval_recall": 0.9705882352941176, + "eval_runtime": 3.1171, + "eval_samples_per_second": 65.446, + "eval_steps_per_second": 8.341, "step": 25 }, { "epoch": 2.980392156862745, - "eval_accuracy": 0.9607843137254902, - "eval_f1": 0.9608937528499771, - "eval_loss": 0.07022649049758911, - "eval_precision": 0.9611995291537668, - "eval_recall": 0.9607843137254902, - "eval_runtime": 3.1392, - "eval_samples_per_second": 64.985, - "eval_steps_per_second": 8.282, + "eval_accuracy": 0.9803921568627451, + "eval_f1": 0.9803921568627451, + "eval_loss": 0.06114682927727699, + "eval_precision": 0.9803921568627451, + "eval_recall": 0.9803921568627451, + "eval_runtime": 3.2163, + "eval_samples_per_second": 63.428, + "eval_steps_per_second": 8.084, "step": 38 }, { "epoch": 4.0, - "eval_accuracy": 0.9705882352941176, - "eval_f1": 0.9707454527863777, - "eval_loss": 0.05151732638478279, - "eval_precision": 0.9716895305130601, - "eval_recall": 0.9705882352941176, - "eval_runtime": 3.1558, - "eval_samples_per_second": 64.644, - "eval_steps_per_second": 8.239, + "eval_accuracy": 0.9852941176470589, + "eval_f1": 0.9852722698274712, + "eval_loss": 0.049169763922691345, + "eval_precision": 0.9853018065952722, + "eval_recall": 0.9852941176470589, + "eval_runtime": 3.1223, + "eval_samples_per_second": 65.336, + "eval_steps_per_second": 8.327, "step": 51 }, { "epoch": 4.9411764705882355, - "eval_accuracy": 0.9754901960784313, - "eval_f1": 0.9754537830457853, - "eval_loss": 0.04813842847943306, - "eval_precision": 0.9754681544268864, - "eval_recall": 0.9754901960784313, - "eval_runtime": 3.2298, - "eval_samples_per_second": 63.162, - "eval_steps_per_second": 8.05, + "eval_accuracy": 0.9803921568627451, + "eval_f1": 0.9804969685242518, + "eval_loss": 0.05879249796271324, + "eval_precision": 0.9813976872800402, + "eval_recall": 0.9803921568627451, + "eval_runtime": 3.1449, + "eval_samples_per_second": 64.868, + "eval_steps_per_second": 8.267, "step": 63 } ], "logging_steps": 500, - "max_steps": 96, + "max_steps": 120, "num_input_tokens_seen": 0, - "num_train_epochs": 8, + "num_train_epochs": 10, "save_steps": 500, "stateful_callbacks": { "EarlyStoppingCallback": { @@ -95,7 +95,7 @@ "attributes": {} } }, - "total_flos": 6.574492142240755e+16, + "total_flos": 6.573703002746803e+16, "train_batch_size": 8, "trial_name": null, "trial_params": null diff --git a/checkpoint-63/training_args.bin b/checkpoint-63/training_args.bin index c9ee3705505943fb60df784486d5cefbd8aa885b..ff5c682679da830384645399a104cd532805ce20 100644 --- a/checkpoint-63/training_args.bin +++ b/checkpoint-63/training_args.bin @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:ecc689edbe3573cb27ca0a4ddd21670a9f672918abacfb53cac5b55b680695a7 +oid sha256:3b77b2a33ca88330b7909aacab81332695e9c9dcf797e7a7757ca31ffcde6f31 size 5240 diff --git a/checkpoint-76/model.safetensors b/checkpoint-76/model.safetensors index 4bcca1631524fb65d1a40c549f362b135de481db..564b70bc93234c5aaf7aa4016ec40dc0418dc83c 100644 --- a/checkpoint-76/model.safetensors +++ b/checkpoint-76/model.safetensors @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:cce24f032e8b31d815639b8572cfd19294eceff51a8bc45bd7ca65a00f26f3c8 +oid sha256:c3d95c32dbd1084212866f7f34d3e27d0b151ce02486333ab299b34fbaf72996 size 94763496 diff --git a/checkpoint-76/optimizer.pt b/checkpoint-76/optimizer.pt index fed388c9400b3e2d109074d518045123f974d63f..d0706f873c12875e0bced75f7154965b61710e12 100644 --- a/checkpoint-76/optimizer.pt +++ b/checkpoint-76/optimizer.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:434bebe7298896598a82dd9ee19ebcf516289a69e3722ee1f3ed5d68cbec8703 +oid sha256:99522ca0bc0a54af9c4c8cbf9642adefc9dfb8f66509991df984da75875dd750 size 189552570 diff --git a/checkpoint-76/rng_state.pth b/checkpoint-76/rng_state.pth index 67eecb5fcf474a86e566537883b7a6d168a65485..e8364ab34a3eb343c908639996df17c999dedb4f 100644 --- a/checkpoint-76/rng_state.pth +++ b/checkpoint-76/rng_state.pth @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:b23c9a094e415e3ed31687390d96e4d360032213140b71a82f2d93753f6eb953 -size 14244 +oid sha256:da6cbce3af0bdcb7fe61b6ecc8ec2df6a6d96cc88f84d7224402bd0e613e62f7 +size 14308 diff --git a/checkpoint-76/scheduler.pt b/checkpoint-76/scheduler.pt index baba98acc162bb6767dec91ac525eba8590da107..a6b94edbba5ad4f2af2f11e0ce38fbebafad4139 100644 --- a/checkpoint-76/scheduler.pt +++ b/checkpoint-76/scheduler.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:eab5147c282d6a9f9aa65e16c32ebff237bd23d33b6eac441cb71ba9dc2610ee +oid sha256:06c9496c56e2fd1b8c8806a5447985117c9db816e9efcb22caec9e9dcd66aa79 size 1064 diff --git a/checkpoint-76/trainer_state.json b/checkpoint-76/trainer_state.json index 982e7d7342796973896ae3c69a01d3c9cf22c6b6..abf96945531fbb48960cdd95aebb341a75dc7c4f 100644 --- a/checkpoint-76/trainer_state.json +++ b/checkpoint-76/trainer_state.json @@ -1,6 +1,6 @@ { - "best_metric": 0.9901960784313726, - "best_model_checkpoint": "distilhubert-finetuned-cry-detector/checkpoint-76", + "best_metric": 0.9852941176470589, + "best_model_checkpoint": "distilhubert-finetuned-cry-detector/checkpoint-51", "epoch": 5.96078431372549, "eval_steps": 500, "global_step": 76, @@ -10,81 +10,81 @@ "log_history": [ { "epoch": 0.9411764705882353, - "eval_accuracy": 0.9558823529411765, - "eval_f1": 0.9558168094824137, - "eval_loss": 0.18768352270126343, - "eval_precision": 0.9558008500901145, - "eval_recall": 0.9558823529411765, - "eval_runtime": 3.1767, - "eval_samples_per_second": 64.218, - "eval_steps_per_second": 8.185, + "eval_accuracy": 0.9362745098039216, + "eval_f1": 0.9365356058160116, + "eval_loss": 0.19313116371631622, + "eval_precision": 0.9372234219385678, + "eval_recall": 0.9362745098039216, + "eval_runtime": 3.0986, + "eval_samples_per_second": 65.836, + "eval_steps_per_second": 8.391, "step": 12 }, { "epoch": 1.9607843137254903, - "eval_accuracy": 0.9607843137254902, - "eval_f1": 0.9603980954872683, - "eval_loss": 0.10839918255805969, - "eval_precision": 0.961865628604383, - "eval_recall": 0.9607843137254902, - "eval_runtime": 3.1396, - "eval_samples_per_second": 64.977, - "eval_steps_per_second": 8.281, + "eval_accuracy": 0.9705882352941176, + "eval_f1": 0.9704025549613785, + "eval_loss": 0.09495726972818375, + "eval_precision": 0.971039759187257, + "eval_recall": 0.9705882352941176, + "eval_runtime": 3.1171, + "eval_samples_per_second": 65.446, + "eval_steps_per_second": 8.341, "step": 25 }, { "epoch": 2.980392156862745, - "eval_accuracy": 0.9607843137254902, - "eval_f1": 0.9608937528499771, - "eval_loss": 0.07022649049758911, - "eval_precision": 0.9611995291537668, - "eval_recall": 0.9607843137254902, - "eval_runtime": 3.1392, - "eval_samples_per_second": 64.985, - "eval_steps_per_second": 8.282, + "eval_accuracy": 0.9803921568627451, + "eval_f1": 0.9803921568627451, + "eval_loss": 0.06114682927727699, + "eval_precision": 0.9803921568627451, + "eval_recall": 0.9803921568627451, + "eval_runtime": 3.2163, + "eval_samples_per_second": 63.428, + "eval_steps_per_second": 8.084, "step": 38 }, { "epoch": 4.0, - "eval_accuracy": 0.9705882352941176, - "eval_f1": 0.9707454527863777, - "eval_loss": 0.05151732638478279, - "eval_precision": 0.9716895305130601, - "eval_recall": 0.9705882352941176, - "eval_runtime": 3.1558, - "eval_samples_per_second": 64.644, - "eval_steps_per_second": 8.239, + "eval_accuracy": 0.9852941176470589, + "eval_f1": 0.9852722698274712, + "eval_loss": 0.049169763922691345, + "eval_precision": 0.9853018065952722, + "eval_recall": 0.9852941176470589, + "eval_runtime": 3.1223, + "eval_samples_per_second": 65.336, + "eval_steps_per_second": 8.327, "step": 51 }, { "epoch": 4.9411764705882355, - "eval_accuracy": 0.9754901960784313, - "eval_f1": 0.9754537830457853, - "eval_loss": 0.04813842847943306, - "eval_precision": 0.9754681544268864, - "eval_recall": 0.9754901960784313, - "eval_runtime": 3.2298, - "eval_samples_per_second": 63.162, - "eval_steps_per_second": 8.05, + "eval_accuracy": 0.9803921568627451, + "eval_f1": 0.9804969685242518, + "eval_loss": 0.05879249796271324, + "eval_precision": 0.9813976872800402, + "eval_recall": 0.9803921568627451, + "eval_runtime": 3.1449, + "eval_samples_per_second": 64.868, + "eval_steps_per_second": 8.267, "step": 63 }, { "epoch": 5.96078431372549, - "eval_accuracy": 0.9901960784313726, - "eval_f1": 0.9901960784313726, - "eval_loss": 0.030143579468131065, - "eval_precision": 0.9901960784313726, - "eval_recall": 0.9901960784313726, - "eval_runtime": 3.162, - "eval_samples_per_second": 64.516, - "eval_steps_per_second": 8.223, + "eval_accuracy": 0.9852941176470589, + "eval_f1": 0.9853150765112866, + "eval_loss": 0.03680075705051422, + "eval_precision": 0.9853868369053048, + "eval_recall": 0.9852941176470589, + "eval_runtime": 3.1698, + "eval_samples_per_second": 64.358, + "eval_steps_per_second": 8.203, "step": 76 } ], "logging_steps": 500, - "max_steps": 96, + "max_steps": 120, "num_input_tokens_seen": 0, - "num_train_epochs": 8, + "num_train_epochs": 10, "save_steps": 500, "stateful_callbacks": { "EarlyStoppingCallback": { @@ -107,7 +107,7 @@ "attributes": {} } }, - "total_flos": 7.888438963974643e+16, + "total_flos": 7.888456247721485e+16, "train_batch_size": 8, "trial_name": null, "trial_params": null diff --git a/checkpoint-76/training_args.bin b/checkpoint-76/training_args.bin index c9ee3705505943fb60df784486d5cefbd8aa885b..ff5c682679da830384645399a104cd532805ce20 100644 --- a/checkpoint-76/training_args.bin +++ b/checkpoint-76/training_args.bin @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:ecc689edbe3573cb27ca0a4ddd21670a9f672918abacfb53cac5b55b680695a7 +oid sha256:3b77b2a33ca88330b7909aacab81332695e9c9dcf797e7a7757ca31ffcde6f31 size 5240 diff --git a/checkpoint-89/model.safetensors b/checkpoint-89/model.safetensors index bdf4cc369faf137859e45570fd7290c6b328dfca..8336a63e05425d332b7fdbd1f78a3fe9a4e0a30a 100644 --- a/checkpoint-89/model.safetensors +++ b/checkpoint-89/model.safetensors @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:cf9411734b368b7ccbccc3c95051c7ead933f0a5a3c72cea52b8628ccbbdadb7 +oid sha256:e3b7a85579e55a511b48446f547a26c169e684642d874407fb134c2f8beac015 size 94763496 diff --git a/checkpoint-89/optimizer.pt b/checkpoint-89/optimizer.pt index 55f6311c5bc18b5caab753cd64df7e2aeae7bfe4..67d065d50f93704ff5140efdc1e67680db7bfaf3 100644 --- a/checkpoint-89/optimizer.pt +++ b/checkpoint-89/optimizer.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:41de97548b1993aacd048293fe48a83333054a8b1f67051c6df33fe07da541a6 +oid sha256:1a059745d8de701c191e384c9cfc794ac7b68844bda8c6b2be4f60fdce33a115 size 189552570 diff --git a/checkpoint-89/rng_state.pth b/checkpoint-89/rng_state.pth index afb8d331b0f08077322655eb25569966c6a90e7f..bc3d64852e423d0b162bf787f9e9f9fd34902e24 100644 --- a/checkpoint-89/rng_state.pth +++ b/checkpoint-89/rng_state.pth @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:006bfe353445f4eca6e7d4c48adcc1abb41147899c1602229ba1904d5b7935e1 -size 14244 +oid sha256:e2028757b5dd60434d78dc91413f1564e35dcfa1ce6e8346503d76007bd2cd53 +size 14308 diff --git a/checkpoint-89/scheduler.pt b/checkpoint-89/scheduler.pt index 43b6de85fc8a84a5bd0aa0ca0cc804633eb0665a..3496f9922a8314a3c0d22483498bd1e317dd4da9 100644 --- a/checkpoint-89/scheduler.pt +++ b/checkpoint-89/scheduler.pt @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:e92a8f2d6b6cee65cf243c474cb4bd497444dd92d0b82440eb4087382fa9c5f7 +oid sha256:6febb509af0d9e00654878e02bf290f49cfe1adb47723e8d3f76dd01f72811c5 size 1064 diff --git a/checkpoint-89/trainer_state.json b/checkpoint-89/trainer_state.json index db1cf9a50880f107ce01e696f89a97733a01bf56..5cd0b278d2840533526c47eb7cb96dfb4c0ad4ed 100644 --- a/checkpoint-89/trainer_state.json +++ b/checkpoint-89/trainer_state.json @@ -1,6 +1,6 @@ { "best_metric": 0.9901960784313726, - "best_model_checkpoint": "distilhubert-finetuned-cry-detector/checkpoint-76", + "best_model_checkpoint": "distilhubert-finetuned-cry-detector/checkpoint-89", "epoch": 6.980392156862745, "eval_steps": 500, "global_step": 89, @@ -10,93 +10,93 @@ "log_history": [ { "epoch": 0.9411764705882353, - "eval_accuracy": 0.9558823529411765, - "eval_f1": 0.9558168094824137, - "eval_loss": 0.18768352270126343, - "eval_precision": 0.9558008500901145, - "eval_recall": 0.9558823529411765, - "eval_runtime": 3.1767, - "eval_samples_per_second": 64.218, - "eval_steps_per_second": 8.185, + "eval_accuracy": 0.9362745098039216, + "eval_f1": 0.9365356058160116, + "eval_loss": 0.19313116371631622, + "eval_precision": 0.9372234219385678, + "eval_recall": 0.9362745098039216, + "eval_runtime": 3.0986, + "eval_samples_per_second": 65.836, + "eval_steps_per_second": 8.391, "step": 12 }, { "epoch": 1.9607843137254903, - "eval_accuracy": 0.9607843137254902, - "eval_f1": 0.9603980954872683, - "eval_loss": 0.10839918255805969, - "eval_precision": 0.961865628604383, - "eval_recall": 0.9607843137254902, - "eval_runtime": 3.1396, - "eval_samples_per_second": 64.977, - "eval_steps_per_second": 8.281, + "eval_accuracy": 0.9705882352941176, + "eval_f1": 0.9704025549613785, + "eval_loss": 0.09495726972818375, + "eval_precision": 0.971039759187257, + "eval_recall": 0.9705882352941176, + "eval_runtime": 3.1171, + "eval_samples_per_second": 65.446, + "eval_steps_per_second": 8.341, "step": 25 }, { "epoch": 2.980392156862745, - "eval_accuracy": 0.9607843137254902, - "eval_f1": 0.9608937528499771, - "eval_loss": 0.07022649049758911, - "eval_precision": 0.9611995291537668, - "eval_recall": 0.9607843137254902, - "eval_runtime": 3.1392, - "eval_samples_per_second": 64.985, - "eval_steps_per_second": 8.282, + "eval_accuracy": 0.9803921568627451, + "eval_f1": 0.9803921568627451, + "eval_loss": 0.06114682927727699, + "eval_precision": 0.9803921568627451, + "eval_recall": 0.9803921568627451, + "eval_runtime": 3.2163, + "eval_samples_per_second": 63.428, + "eval_steps_per_second": 8.084, "step": 38 }, { "epoch": 4.0, - "eval_accuracy": 0.9705882352941176, - "eval_f1": 0.9707454527863777, - "eval_loss": 0.05151732638478279, - "eval_precision": 0.9716895305130601, - "eval_recall": 0.9705882352941176, - "eval_runtime": 3.1558, - "eval_samples_per_second": 64.644, - "eval_steps_per_second": 8.239, + "eval_accuracy": 0.9852941176470589, + "eval_f1": 0.9852722698274712, + "eval_loss": 0.049169763922691345, + "eval_precision": 0.9853018065952722, + "eval_recall": 0.9852941176470589, + "eval_runtime": 3.1223, + "eval_samples_per_second": 65.336, + "eval_steps_per_second": 8.327, "step": 51 }, { "epoch": 4.9411764705882355, - "eval_accuracy": 0.9754901960784313, - "eval_f1": 0.9754537830457853, - "eval_loss": 0.04813842847943306, - "eval_precision": 0.9754681544268864, - "eval_recall": 0.9754901960784313, - "eval_runtime": 3.2298, - "eval_samples_per_second": 63.162, - "eval_steps_per_second": 8.05, + "eval_accuracy": 0.9803921568627451, + "eval_f1": 0.9804969685242518, + "eval_loss": 0.05879249796271324, + "eval_precision": 0.9813976872800402, + "eval_recall": 0.9803921568627451, + "eval_runtime": 3.1449, + "eval_samples_per_second": 64.868, + "eval_steps_per_second": 8.267, "step": 63 }, { "epoch": 5.96078431372549, - "eval_accuracy": 0.9901960784313726, - "eval_f1": 0.9901960784313726, - "eval_loss": 0.030143579468131065, - "eval_precision": 0.9901960784313726, - "eval_recall": 0.9901960784313726, - "eval_runtime": 3.162, - "eval_samples_per_second": 64.516, - "eval_steps_per_second": 8.223, + "eval_accuracy": 0.9852941176470589, + "eval_f1": 0.9853150765112866, + "eval_loss": 0.03680075705051422, + "eval_precision": 0.9853868369053048, + "eval_recall": 0.9852941176470589, + "eval_runtime": 3.1698, + "eval_samples_per_second": 64.358, + "eval_steps_per_second": 8.203, "step": 76 }, { "epoch": 6.980392156862745, "eval_accuracy": 0.9901960784313726, - "eval_f1": 0.9901960784313726, - "eval_loss": 0.03071291744709015, - "eval_precision": 0.9901960784313726, + "eval_f1": 0.9901663478316141, + "eval_loss": 0.0382298119366169, + "eval_precision": 0.9903446226975638, "eval_recall": 0.9901960784313726, - "eval_runtime": 3.1655, - "eval_samples_per_second": 64.445, - "eval_steps_per_second": 8.214, + "eval_runtime": 3.1376, + "eval_samples_per_second": 65.019, + "eval_steps_per_second": 8.287, "step": 89 } ], "logging_steps": 500, - "max_steps": 96, + "max_steps": 120, "num_input_tokens_seen": 0, - "num_train_epochs": 8, + "num_train_epochs": 10, "save_steps": 500, "stateful_callbacks": { "EarlyStoppingCallback": { @@ -119,7 +119,7 @@ "attributes": {} } }, - "total_flos": 9.20486736788663e+16, + "total_flos": 9.205556443583078e+16, "train_batch_size": 8, "trial_name": null, "trial_params": null diff --git a/checkpoint-89/training_args.bin b/checkpoint-89/training_args.bin index c9ee3705505943fb60df784486d5cefbd8aa885b..ff5c682679da830384645399a104cd532805ce20 100644 --- a/checkpoint-89/training_args.bin +++ b/checkpoint-89/training_args.bin @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:ecc689edbe3573cb27ca0a4ddd21670a9f672918abacfb53cac5b55b680695a7 +oid sha256:3b77b2a33ca88330b7909aacab81332695e9c9dcf797e7a7757ca31ffcde6f31 size 5240 diff --git a/model.safetensors b/model.safetensors index 99bb8b757cc82dc894263c1ed673d5ba4340cbe6..b63e1b217d7bda6f42a6ac5fe5cc72b618e9bffe 100644 --- a/model.safetensors +++ b/model.safetensors @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:ed24ddc80fbfa0595409861cddf08e518561fbdd7c4461e5edac2ef1da033508 +oid sha256:098c90d2cba195ac20033090c5954f7ef596abcdeebba724cb1365151033b05e size 94763496 diff --git a/runs/Sep01_13-01-56_ubumarcos/events.out.tfevents.1725188518.ubumarcos b/runs/Sep01_13-01-56_ubumarcos/events.out.tfevents.1725188518.ubumarcos new file mode 100644 index 0000000000000000000000000000000000000000..6c82e68e760ef3f7730d8e8db31b48f7caadd671 --- /dev/null +++ b/runs/Sep01_13-01-56_ubumarcos/events.out.tfevents.1725188518.ubumarcos @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:47699197c72dcec085f85612f80b753c7cc8dbe0771b4fc7e232ef7fbf4ed4c0 +size 10789 diff --git a/training_args.bin b/training_args.bin index 17bea61543f565c72b3b723abec4a0277f862e92..ff5c682679da830384645399a104cd532805ce20 100644 --- a/training_args.bin +++ b/training_args.bin @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:a1c27b4ecab2c1ddc610f4ce2ead133759bcabe788a27ce90b2b23560c4e6ad3 +oid sha256:3b77b2a33ca88330b7909aacab81332695e9c9dcf797e7a7757ca31ffcde6f31 size 5240