Delete checkpoint-330

checkpoint-330/config.json

@@ -1,119 +0,0 @@
-{
-  "architectures": [
-    "ModernBertForTokenClassification"
-  ],
-  "attention_bias": false,
-  "attention_dropout": 0.0,
-  "bos_token_id": 50281,
-  "classifier_activation": "gelu",
-  "classifier_bias": false,
-  "classifier_dropout": 0.0,
-  "classifier_pooling": "mean",
-  "cls_token_id": 50281,
-  "decoder_bias": true,
-  "deterministic_flash_attn": false,
-  "embedding_dropout": 0.0,
-  "eos_token_id": 50282,
-  "global_attn_every_n_layers": 3,
-  "global_rope_theta": 160000.0,
-  "gradient_checkpointing": false,
-  "hidden_activation": "gelu",
-  "hidden_size": 768,
-  "id2label": {
-    "0": "O",
-    "1": "B-AGE",
-    "2": "I-AGE",
-    "3": "B-CREDIT_CARD",
-    "4": "I-CREDIT_CARD",
-    "5": "B-DATE_TIME",
-    "6": "I-DATE_TIME",
-    "7": "B-DOMAIN_NAME",
-    "8": "I-DOMAIN_NAME",
-    "9": "B-EMAIL_ADDRESS",
-    "10": "I-EMAIL_ADDRESS",
-    "11": "B-GPE",
-    "12": "I-GPE",
-    "13": "B-IBAN_CODE",
-    "14": "I-IBAN_CODE",
-    "15": "B-IP_ADDRESS",
-    "16": "I-IP_ADDRESS",
-    "17": "B-NRP",
-    "18": "I-NRP",
-    "19": "B-ORGANIZATION",
-    "20": "I-ORGANIZATION",
-    "21": "B-PERSON",
-    "22": "I-PERSON",
-    "23": "B-PHONE_NUMBER",
-    "24": "I-PHONE_NUMBER",
-    "25": "B-STREET_ADDRESS",
-    "26": "I-STREET_ADDRESS",
-    "27": "B-TITLE",
-    "28": "I-TITLE",
-    "29": "B-US_DRIVER_LICENSE",
-    "30": "I-US_DRIVER_LICENSE",
-    "31": "B-US_SSN",
-    "32": "I-US_SSN",
-    "33": "B-ZIP_CODE",
-    "34": "I-ZIP_CODE"
-  },
-  "initializer_cutoff_factor": 2.0,
-  "initializer_range": 0.02,
-  "intermediate_size": 1152,
-  "label2id": {
-    "B-AGE": 1,
-    "B-CREDIT_CARD": 3,
-    "B-DATE_TIME": 5,
-    "B-DOMAIN_NAME": 7,
-    "B-EMAIL_ADDRESS": 9,
-    "B-GPE": 11,
-    "B-IBAN_CODE": 13,
-    "B-IP_ADDRESS": 15,
-    "B-NRP": 17,
-    "B-ORGANIZATION": 19,
-    "B-PERSON": 21,
-    "B-PHONE_NUMBER": 23,
-    "B-STREET_ADDRESS": 25,
-    "B-TITLE": 27,
-    "B-US_DRIVER_LICENSE": 29,
-    "B-US_SSN": 31,
-    "B-ZIP_CODE": 33,
-    "I-AGE": 2,
-    "I-CREDIT_CARD": 4,
-    "I-DATE_TIME": 6,
-    "I-DOMAIN_NAME": 8,
-    "I-EMAIL_ADDRESS": 10,
-    "I-GPE": 12,
-    "I-IBAN_CODE": 14,
-    "I-IP_ADDRESS": 16,
-    "I-NRP": 18,
-    "I-ORGANIZATION": 20,
-    "I-PERSON": 22,
-    "I-PHONE_NUMBER": 24,
-    "I-STREET_ADDRESS": 26,
-    "I-TITLE": 28,
-    "I-US_DRIVER_LICENSE": 30,
-    "I-US_SSN": 32,
-    "I-ZIP_CODE": 34,
-    "O": 0
-  },
-  "layer_norm_eps": 1e-05,
-  "local_attention": 128,
-  "local_rope_theta": 10000.0,
-  "max_position_embeddings": 8192,
-  "mlp_bias": false,
-  "mlp_dropout": 0.0,
-  "model_type": "modernbert",
-  "norm_bias": false,
-  "norm_eps": 1e-05,
-  "num_attention_heads": 12,
-  "num_hidden_layers": 22,
-  "pad_token_id": 50283,
-  "position_embedding_type": "absolute",
-  "repad_logits_with_grad": false,
-  "sep_token_id": 50282,
-  "sparse_pred_ignore_index": -100,
-  "sparse_prediction": false,
-  "torch_dtype": "float32",
-  "transformers_version": "4.55.2",
-  "vocab_size": 50368
-}

checkpoint-330/model.safetensors

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:cb6ec1698bff6906a665266266f3f8b2e743b6619236b6e7c4fa74a76f0715ca
-size 598541300

checkpoint-330/optimizer.pt

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:175a3fdb326dd0f591e2398306ac7a215399ef4d9545e6aa4c3b44a8353298a8
-size 1197172747

checkpoint-330/rng_state.pth

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:ed3753ab7977739b8eda494dd72defae5750f7283141b11a8f562160ba4c1a23
-size 14645

checkpoint-330/scheduler.pt

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:91ca43ad54404cb852f61f3179fa0d50369b7bed2a3d846ad14d8537ea3c12a9
-size 1465

checkpoint-330/trainer_state.json

@@ -1,175 +0,0 @@
-{
-  "best_global_step": 330,
-  "best_metric": 0.9632840387557369,
-  "best_model_checkpoint": "pii_classifier_modernbert-base_presidio_token_model/checkpoint-330",
-  "epoch": 10.0,
-  "eval_steps": 500,
-  "global_step": 330,
-  "is_hyper_param_search": false,
-  "is_local_process_zero": true,
-  "is_world_process_zero": true,
-  "log_history": [
-    {
-      "epoch": 1.0,
-      "eval_accuracy": 0.5799762026177121,
-      "eval_f1": 0.4695039711201284,
-      "eval_loss": 1.893957495689392,
-      "eval_precision": 0.4935126787661159,
-      "eval_recall": 0.5799762026177121,
-      "eval_runtime": 23.603,
-      "eval_samples_per_second": 9.533,
-      "eval_steps_per_second": 0.636,
-      "step": 33
-    },
-    {
-      "epoch": 2.0,
-      "eval_accuracy": 0.7533571307156213,
-      "eval_f1": 0.7111688388751078,
-      "eval_loss": 0.9364706873893738,
-      "eval_precision": 0.7083445547100621,
-      "eval_recall": 0.7533571307156213,
-      "eval_runtime": 23.6426,
-      "eval_samples_per_second": 9.517,
-      "eval_steps_per_second": 0.634,
-      "step": 66
-    },
-    {
-      "epoch": 3.0,
-      "eval_accuracy": 0.8733639299677035,
-      "eval_f1": 0.8472559039016181,
-      "eval_loss": 0.4803955852985382,
-      "eval_precision": 0.8461851394388445,
-      "eval_recall": 0.8733639299677035,
-      "eval_runtime": 23.7467,
-      "eval_samples_per_second": 9.475,
-      "eval_steps_per_second": 0.632,
-      "step": 99
-    },
-    {
-      "epoch": 3.0303030303030303,
-      "grad_norm": 3.357513427734375,
-      "learning_rate": 8.83022221559489e-06,
-      "loss": 1.7939,
-      "step": 100
-    },
-    {
-      "epoch": 4.0,
-      "eval_accuracy": 0.9144994050654428,
-      "eval_f1": 0.9046430604528387,
-      "eval_loss": 0.30626583099365234,
-      "eval_precision": 0.9007848165265108,
-      "eval_recall": 0.9144994050654428,
-      "eval_runtime": 23.9053,
-      "eval_samples_per_second": 9.412,
-      "eval_steps_per_second": 0.627,
-      "step": 132
-    },
-    {
-      "epoch": 5.0,
-      "eval_accuracy": 0.9464558898521163,
-      "eval_f1": 0.941120594338562,
-      "eval_loss": 0.20138122141361237,
-      "eval_precision": 0.9377487874133417,
-      "eval_recall": 0.9464558898521163,
-      "eval_runtime": 23.6064,
-      "eval_samples_per_second": 9.531,
-      "eval_steps_per_second": 0.635,
-      "step": 165
-    },
-    {
-      "epoch": 6.0,
-      "eval_accuracy": 0.9552949175590685,
-      "eval_f1": 0.9524039016337489,
-      "eval_loss": 0.16908280551433563,
-      "eval_precision": 0.9532284447896412,
-      "eval_recall": 0.9552949175590685,
-      "eval_runtime": 23.516,
-      "eval_samples_per_second": 9.568,
-      "eval_steps_per_second": 0.638,
-      "step": 198
-    },
-    {
-      "epoch": 6.0606060606060606,
-      "grad_norm": 1.2425391674041748,
-      "learning_rate": 4.079723389713899e-06,
-      "loss": 0.1942,
-      "step": 200
-    },
-    {
-      "epoch": 7.0,
-      "eval_accuracy": 0.960224375318715,
-      "eval_f1": 0.9576892292455826,
-      "eval_loss": 0.1495475322008133,
-      "eval_precision": 0.9587472552812538,
-      "eval_recall": 0.960224375318715,
-      "eval_runtime": 23.9106,
-      "eval_samples_per_second": 9.41,
-      "eval_steps_per_second": 0.627,
-      "step": 231
-    },
-    {
-      "epoch": 8.0,
-      "eval_accuracy": 0.9631140574536801,
-      "eval_f1": 0.9613609203737836,
-      "eval_loss": 0.1422007530927658,
-      "eval_precision": 0.9626175146629873,
-      "eval_recall": 0.9631140574536801,
-      "eval_runtime": 23.6358,
-      "eval_samples_per_second": 9.519,
-      "eval_steps_per_second": 0.635,
-      "step": 264
-    },
-    {
-      "epoch": 9.0,
-      "eval_accuracy": 0.9629440761516234,
-      "eval_f1": 0.9614999857063931,
-      "eval_loss": 0.13786566257476807,
-      "eval_precision": 0.9629680400789847,
-      "eval_recall": 0.9629440761516234,
-      "eval_runtime": 23.9025,
-      "eval_samples_per_second": 9.413,
-      "eval_steps_per_second": 0.628,
-      "step": 297
-    },
-    {
-      "epoch": 9.090909090909092,
-      "grad_norm": 1.0872043371200562,
-      "learning_rate": 2.664129206497479e-07,
-      "loss": 0.0463,
-      "step": 300
-    },
-    {
-      "epoch": 10.0,
-      "eval_accuracy": 0.9632840387557369,
-      "eval_f1": 0.961914649184486,
-      "eval_loss": 0.13768431544303894,
-      "eval_precision": 0.9633718277416998,
-      "eval_recall": 0.9632840387557369,
-      "eval_runtime": 23.7603,
-      "eval_samples_per_second": 9.47,
-      "eval_steps_per_second": 0.631,
-      "step": 330
-    }
-  ],
-  "logging_steps": 100,
-  "max_steps": 330,
-  "num_input_tokens_seen": 0,
-  "num_train_epochs": 10,
-  "save_steps": 500,
-  "stateful_callbacks": {
-    "TrainerControl": {
-      "args": {
-        "should_epoch_stop": false,
-        "should_evaluate": false,
-        "should_log": false,
-        "should_save": true,
-        "should_training_stop": true
-      },
-      "attributes": {}
-    }
-  },
-  "total_flos": 2.863020754944e+16,
-  "train_batch_size": 16,
-  "trial_name": null,
-  "trial_params": null
-}

checkpoint-330/training_args.bin

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:315e6b72e779eb2321d1291722c827013c190ff330ee5e8ee58c3e70fe5772a5
-size 5777