Upload folder using huggingface_hub

training_checkpoints/checkpoint-1344/added_tokens.json

@@ -0,0 +1,3 @@
+{
+  "<pad>": 30522
+}

training_checkpoints/checkpoint-1344/config.json

@@ -0,0 +1,41 @@
+{
+  "_name_or_path": "law-ai/InLegalBERT",
+  "architectures": [
+    "BertForSequenceClassification"
+  ],
+  "attention_probs_dropout_prob": 0.1,
+  "bos_token_id": 0,
+  "classifier_dropout": null,
+  "eos_token_ids": 0,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 768,
+  "id2label": {
+    "0": "positive",
+    "1": "negative",
+    "2": "neutral",
+    "3": "rel_by_party"
+  },
+  "initializer_range": 0.02,
+  "intermediate_size": 3072,
+  "label2id": {
+    "negative": 1,
+    "neutral": 2,
+    "positive": 0,
+    "rel_by_party": 3
+  },
+  "layer_norm_eps": 1e-12,
+  "max_position_embeddings": 512,
+  "model_type": "bert",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 12,
+  "output_past": true,
+  "pad_token_id": 30522,
+  "position_embedding_type": "absolute",
+  "problem_type": "single_label_classification",
+  "torch_dtype": "float32",
+  "transformers_version": "4.39.3",
+  "type_vocab_size": 2,
+  "use_cache": true,
+  "vocab_size": 30523
+}

training_checkpoints/checkpoint-1344/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1f77bcb241ce3e3d0248f4ccdfe28819129e26f1feef08fb2b05fa0ad661ce47
+size 437967872

training_checkpoints/checkpoint-1344/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:69240b464965fc6f5ddc223703145d0ce21d0acb62d436f26afb79f94f816c0b
+size 876056826

training_checkpoints/checkpoint-1344/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a81cd86c86c94d2bbbdc677771a885c9ab0fc780da7e31b40ece3ef68fafd74b
+size 14244

training_checkpoints/checkpoint-1344/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a6d586e83ca8e6f1b6218ee8f89eb3cb986c75a98d649cc24f6a2751d3843bca
+size 1064

training_checkpoints/checkpoint-1344/special_tokens_map.json

@@ -0,0 +1,13 @@
+{
+  "cls_token": "[CLS]",
+  "mask_token": "[MASK]",
+  "pad_token": {
+    "content": "<pad>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "sep_token": "[SEP]",
+  "unk_token": "[UNK]"
+}

training_checkpoints/checkpoint-1344/tokenizer_config.json

@@ -0,0 +1,65 @@
+{
+  "added_tokens_decoder": {
+    "0": {
+      "content": "[PAD]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "100": {
+      "content": "[UNK]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "101": {
+      "content": "[CLS]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "102": {
+      "content": "[SEP]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "103": {
+      "content": "[MASK]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "30522": {
+      "content": "<pad>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "clean_up_tokenization_spaces": true,
+  "cls_token": "[CLS]",
+  "do_basic_tokenize": true,
+  "do_lower_case": true,
+  "mask_token": "[MASK]",
+  "model_max_length": 512,
+  "never_split": null,
+  "pad_token": "<pad>",
+  "sep_token": "[SEP]",
+  "strip_accents": null,
+  "tokenize_chinese_chars": true,
+  "tokenizer_class": "BertTokenizer",
+  "unk_token": "[UNK]"
+}

training_checkpoints/checkpoint-1344/trainer_state.json

@@ -0,0 +1,167 @@
+{
+  "best_metric": 0.6690246389304302,
+  "best_model_checkpoint": "case-analysis-InLegalBERT/checkpoint-1344",
+  "epoch": 6.0,
+  "eval_steps": 500,
+  "global_step": 1344,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 1.0,
+      "eval_accuracy": 0.8017817371937639,
+      "eval_f1_macro": 0.5936102522033613,
+      "eval_f1_micro": 0.8017817371937639,
+      "eval_f1_weighted": 0.7819923339877444,
+      "eval_loss": 0.6545999050140381,
+      "eval_macro_fpr": 0.09781883735656886,
+      "eval_macro_sensitivity": 0.6106003826592061,
+      "eval_macro_specificity": 0.9112420642353884,
+      "eval_precision": 0.763240968439436,
+      "eval_precision_macro": 0.5776616905319997,
+      "eval_recall": 0.8017817371937639,
+      "eval_recall_macro": 0.6106003826592061,
+      "eval_runtime": 9.2925,
+      "eval_samples_per_second": 48.318,
+      "eval_steps_per_second": 6.134,
+      "eval_weighted_fpr": 0.07613344739093242,
+      "eval_weighted_sensitivity": 0.8017817371937639,
+      "eval_weighted_specificity": 0.84318651974779,
+      "step": 224
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": 0.8129175946547884,
+      "eval_f1_macro": 0.5995597405020452,
+      "eval_f1_micro": 0.8129175946547884,
+      "eval_f1_weighted": 0.7925821933419933,
+      "eval_loss": 0.6830806732177734,
+      "eval_macro_fpr": 0.0922965769071398,
+      "eval_macro_sensitivity": 0.615426637485461,
+      "eval_macro_specificity": 0.9170902440637451,
+      "eval_precision": 0.7732439444566204,
+      "eval_precision_macro": 0.5844969763606377,
+      "eval_recall": 0.8129175946547884,
+      "eval_recall_macro": 0.615426637485461,
+      "eval_runtime": 36.5377,
+      "eval_samples_per_second": 12.289,
+      "eval_steps_per_second": 1.56,
+      "eval_weighted_fpr": 0.07124681933842239,
+      "eval_weighted_sensitivity": 0.8129175946547884,
+      "eval_weighted_specificity": 0.8554433816001921,
+      "step": 448
+    },
+    {
+      "epoch": 2.23,
+      "grad_norm": 6.910687446594238,
+      "learning_rate": 4.6302083333333335e-05,
+      "loss": 0.607,
+      "step": 500
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": 0.8262806236080178,
+      "eval_f1_macro": 0.6362135046138226,
+      "eval_f1_micro": 0.8262806236080178,
+      "eval_f1_weighted": 0.8105096248986687,
+      "eval_loss": 0.7626442909240723,
+      "eval_macro_fpr": 0.0885057964533289,
+      "eval_macro_sensitivity": 0.6341414944356121,
+      "eval_macro_specificity": 0.9181782637863818,
+      "eval_precision": 0.8060163824047696,
+      "eval_precision_macro": 0.6772816989326722,
+      "eval_recall": 0.8262806236080178,
+      "eval_recall_macro": 0.6341414944356121,
+      "eval_runtime": 35.6726,
+      "eval_samples_per_second": 12.587,
+      "eval_steps_per_second": 1.598,
+      "eval_weighted_fpr": 0.0654911838790932,
+      "eval_weighted_sensitivity": 0.8262806236080178,
+      "eval_weighted_specificity": 0.8464324315375095,
+      "step": 672
+    },
+    {
+      "epoch": 4.0,
+      "eval_accuracy": 0.8084632516703786,
+      "eval_f1_macro": 0.6313839669673806,
+      "eval_f1_micro": 0.8084632516703786,
+      "eval_f1_weighted": 0.8017481865377005,
+      "eval_loss": 0.7839008569717407,
+      "eval_macro_fpr": 0.08961602877899306,
+      "eval_macro_sensitivity": 0.630579873226932,
+      "eval_macro_specificity": 0.9209618973154744,
+      "eval_precision": 0.7990879202460494,
+      "eval_precision_macro": 0.6390850934968582,
+      "eval_recall": 0.8084632516703786,
+      "eval_recall_macro": 0.630579873226932,
+      "eval_runtime": 33.6763,
+      "eval_samples_per_second": 13.333,
+      "eval_steps_per_second": 1.693,
+      "eval_weighted_fpr": 0.07319148936170213,
+      "eval_weighted_sensitivity": 0.8084632516703786,
+      "eval_weighted_specificity": 0.8753843375915193,
+      "step": 896
+    },
+    {
+      "epoch": 4.46,
+      "grad_norm": 0.3201627731323242,
+      "learning_rate": 4.258184523809524e-05,
+      "loss": 0.316,
+      "step": 1000
+    },
+    {
+      "epoch": 5.0,
+      "eval_accuracy": 0.8262806236080178,
+      "eval_f1_macro": 0.6513945578231293,
+      "eval_f1_micro": 0.8262806236080178,
+      "eval_f1_weighted": 0.8161065406117904,
+      "eval_loss": 0.9380576610565186,
+      "eval_macro_fpr": 0.08224137352531583,
+      "eval_macro_sensitivity": 0.6573161205514146,
+      "eval_macro_specificity": 0.9260657452373569,
+      "eval_precision": 0.8127349298285333,
+      "eval_precision_macro": 0.668810150869688,
+      "eval_recall": 0.8262806236080178,
+      "eval_recall_macro": 0.6573161205514146,
+      "eval_runtime": 34.3275,
+      "eval_samples_per_second": 13.08,
+      "eval_steps_per_second": 1.66,
+      "eval_weighted_fpr": 0.0654911838790932,
+      "eval_weighted_sensitivity": 0.8262806236080178,
+      "eval_weighted_specificity": 0.8779823573414097,
+      "step": 1120
+    },
+    {
+      "epoch": 6.0,
+      "eval_accuracy": 0.821826280623608,
+      "eval_f1_macro": 0.6690246389304302,
+      "eval_f1_micro": 0.821826280623608,
+      "eval_f1_weighted": 0.8159237291984361,
+      "eval_loss": 1.0433590412139893,
+      "eval_macro_fpr": 0.08967976675610558,
+      "eval_macro_sensitivity": 0.6533037047742931,
+      "eval_macro_specificity": 0.9186687903009794,
+      "eval_precision": 0.8145202780918083,
+      "eval_precision_macro": 0.6906556062141168,
+      "eval_recall": 0.821826280623608,
+      "eval_recall_macro": 0.6533037047742931,
+      "eval_runtime": 43.316,
+      "eval_samples_per_second": 10.366,
+      "eval_steps_per_second": 1.316,
+      "eval_weighted_fpr": 0.06739679865206402,
+      "eval_weighted_sensitivity": 0.821826280623608,
+      "eval_weighted_specificity": 0.8528488805803094,
+      "step": 1344
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 6720,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 30,
+  "save_steps": 500,
+  "total_flos": 2829020867592192.0,
+  "train_batch_size": 8,
+  "trial_name": null,
+  "trial_params": null
+}

training_checkpoints/checkpoint-1344/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4168c8c84941deb81d05fe2d714c7e20c482e1e8b4900f8d6b307d6f8c3828dd
+size 4984