End of training

README.md

@@ -3,9 +3,10 @@ library_name: transformers
 license: apache-2.0
 base_model: PlanTL-GOB-ES/bsc-bio-ehr-es
 tags:
+- token-classification
 - generated_from_trainer
 datasets:
-- combined-train-distemist-dev-85-ner
+- Rodrigo1771/combined-train-distemist-dev-85-ner
 metrics:
 - precision
 - recall
@@ -18,24 +19,24 @@ model-index:
       name: Token Classification
       type: token-classification
     dataset:
-      name: combined-train-distemist-dev-85-ner
-      type: combined-train-distemist-dev-85-ner
+      name: Rodrigo1771/combined-train-distemist-dev-85-ner
+      type: Rodrigo1771/combined-train-distemist-dev-85-ner
       config: CombinedTrainDisTEMISTDevNER
       split: validation
       args: CombinedTrainDisTEMISTDevNER
     metrics:
     - name: Precision
       type: precision
-      value: 0.31162999550965426
+      value: 0.3152508603513856
     - name: Recall
       type: recall
-      value: 0.8118858212447356
+      value: 0.8144595226953674
     - name: F1
       type: f1
-      value: 0.45038613797131544
+      value: 0.45455732567249935
     - name: Accuracy
       type: accuracy
-      value: 0.8533304955579661
+      value: 0.8564886649182308
 ---
 
 <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -43,13 +44,13 @@ should probably proofread and complete it, then remove this comment. -->
 
 # output
 
-This model is a fine-tuned version of [PlanTL-GOB-ES/bsc-bio-ehr-es](https://huggingface.co/PlanTL-GOB-ES/bsc-bio-ehr-es) on the combined-train-distemist-dev-85-ner dataset.
+This model is a fine-tuned version of [PlanTL-GOB-ES/bsc-bio-ehr-es](https://huggingface.co/PlanTL-GOB-ES/bsc-bio-ehr-es) on the Rodrigo1771/combined-train-distemist-dev-85-ner dataset.
 It achieves the following results on the evaluation set:
-- Loss: 1.0023
-- Precision: 0.3116
-- Recall: 0.8119
-- F1: 0.4504
-- Accuracy: 0.8533
+- Loss: 0.7006
+- Precision: 0.3153
+- Recall: 0.8145
+- F1: 0.4546
+- Accuracy: 0.8565
 
 ## Model description
 

all_results.json

@@ -1,26 +1,26 @@
 {
     "epoch": 10.0,
-    "eval_accuracy": 0.9989426998228679,
-    "eval_f1": 0.9491525423728814,
-    "eval_loss": 0.004777050111442804,
-    "eval_precision": 0.9461187214611873,
-    "eval_recall": 0.9522058823529411,
-    "eval_runtime": 13.9476,
+    "eval_accuracy": 0.8564886649182308,
+    "eval_f1": 0.45455732567249935,
+    "eval_loss": 0.7005925178527832,
+    "eval_precision": 0.3152508603513856,
+    "eval_recall": 0.8144595226953674,
+    "eval_runtime": 14.2934,
     "eval_samples": 6810,
-    "eval_samples_per_second": 488.256,
-    "eval_steps_per_second": 61.086,
-    "predict_accuracy": 0.9987478324070453,
-    "predict_f1": 0.9243073407597828,
-    "predict_loss": 0.005894536152482033,
-    "predict_precision": 0.9069506726457399,
-    "predict_recall": 0.9423412929528246,
-    "predict_runtime": 28.2179,
-    "predict_samples_per_second": 517.899,
-    "predict_steps_per_second": 64.746,
-    "total_flos": 1.4714840952259542e+16,
-    "train_loss": 0.002772659832779558,
-    "train_runtime": 1349.0548,
-    "train_samples": 29797,
-    "train_samples_per_second": 220.873,
-    "train_steps_per_second": 3.454
+    "eval_samples_per_second": 476.445,
+    "eval_steps_per_second": 59.608,
+    "predict_accuracy": 0.9437946603149774,
+    "predict_f1": 0.6365870441364396,
+    "predict_loss": 0.22766011953353882,
+    "predict_precision": 0.519576379974326,
+    "predict_recall": 0.8216188784572444,
+    "predict_runtime": 29.1056,
+    "predict_samples_per_second": 502.103,
+    "predict_steps_per_second": 62.771,
+    "total_flos": 1.7176580067661056e+16,
+    "train_loss": 0.0812657987344287,
+    "train_runtime": 1549.44,
+    "train_samples": 34604,
+    "train_samples_per_second": 223.332,
+    "train_steps_per_second": 3.492
 }

eval_results.json

@@ -1,12 +1,12 @@
 {
     "epoch": 10.0,
-    "eval_accuracy": 0.9989426998228679,
-    "eval_f1": 0.9491525423728814,
-    "eval_loss": 0.004777050111442804,
-    "eval_precision": 0.9461187214611873,
-    "eval_recall": 0.9522058823529411,
-    "eval_runtime": 13.9476,
+    "eval_accuracy": 0.8564886649182308,
+    "eval_f1": 0.45455732567249935,
+    "eval_loss": 0.7005925178527832,
+    "eval_precision": 0.3152508603513856,
+    "eval_recall": 0.8144595226953674,
+    "eval_runtime": 14.2934,
     "eval_samples": 6810,
-    "eval_samples_per_second": 488.256,
-    "eval_steps_per_second": 61.086
+    "eval_samples_per_second": 476.445,
+    "eval_steps_per_second": 59.608
 }

predict_results.json

@@ -1,10 +1,10 @@
 {
-    "predict_accuracy": 0.9987478324070453,
-    "predict_f1": 0.9243073407597828,
-    "predict_loss": 0.005894536152482033,
-    "predict_precision": 0.9069506726457399,
-    "predict_recall": 0.9423412929528246,
-    "predict_runtime": 28.2179,
-    "predict_samples_per_second": 517.899,
-    "predict_steps_per_second": 64.746
+    "predict_accuracy": 0.9437946603149774,
+    "predict_f1": 0.6365870441364396,
+    "predict_loss": 0.22766011953353882,
+    "predict_precision": 0.519576379974326,
+    "predict_recall": 0.8216188784572444,
+    "predict_runtime": 29.1056,
+    "predict_samples_per_second": 502.103,
+    "predict_steps_per_second": 62.771
 }

tb/events.out.tfevents.1725581040.2a66098fac87.15776.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:35dfb83cada8fb681a65313600c2857481bd7212900b1573ca8d6ce9b64470bb
+size 560

train.log

@@ -1495,3 +1495,53 @@ Training completed. Do not forget to share your model on huggingface.co/models =
 {'eval_loss': 1.0022608041763306, 'eval_precision': 0.31162999550965426, 'eval_recall': 0.8118858212447356, 'eval_f1': 0.45038613797131544, 'eval_accuracy': 0.8533304955579661, 'eval_runtime': 14.4559, 'eval_samples_per_second': 471.089, 'eval_steps_per_second': 58.938, 'epoch': 10.0}
 {'train_runtime': 1549.44, 'train_samples_per_second': 223.332, 'train_steps_per_second': 3.492, 'train_loss': 0.0812657987344287, 'epoch': 10.0}
 
+***** train metrics *****
+  epoch                    =       10.0
+  total_flos               = 15996936GF
+  train_loss               =     0.0813
+  train_runtime            = 0:25:49.44
+  train_samples            =      34604
+  train_samples_per_second =    223.332
+  train_steps_per_second   =      3.492
+09/06/2024 00:03:46 - INFO - __main__ -   *** Evaluate ***
+[INFO|trainer.py:811] 2024-09-06 00:03:46,664 >> The following columns in the evaluation set don't have a corresponding argument in `RobertaForTokenClassification.forward` and have been ignored: id, ner_tags, tokens. If id, ner_tags, tokens are not expected by `RobertaForTokenClassification.forward`,  you can safely ignore this message.
+[INFO|trainer.py:3819] 2024-09-06 00:03:46,667 >> 
+***** Running Evaluation *****
+[INFO|trainer.py:3821] 2024-09-06 00:03:46,667 >>   Num examples = 6810
+[INFO|trainer.py:3824] 2024-09-06 00:03:46,667 >>   Batch size = 8
+
  0%|          | 0/852 [00:00<?, ?it/s]
  1%|          | 10/852 [00:00<00:09, 92.77it/s]
  2%|▏         | 20/852 [00:00<00:10, 80.16it/s]
  3%|▎         | 29/852 [00:00<00:10, 79.89it/s]
  4%|▍         | 38/852 [00:00<00:10, 80.51it/s]
  6%|▌         | 47/852 [00:00<00:09, 81.99it/s]
  7%|▋         | 56/852 [00:00<00:09, 82.94it/s]
  8%|▊         | 65/852 [00:00<00:09, 82.00it/s]
  9%|▊         | 74/852 [00:00<00:09, 80.66it/s]
 10%|▉         | 83/852 [00:01<00:09, 80.65it/s]
 11%|█         | 92/852 [00:01<00:09, 80.64it/s]
 12%|█▏        | 101/852 [00:01<00:09, 79.84it/s]
 13%|█▎        | 109/852 [00:01<00:09, 79.60it/s]
 14%|█▍        | 118/852 [00:01<00:09, 80.66it/s]
 15%|█▍        | 127/852 [00:01<00:09, 77.73it/s]
 16%|█▌        | 136/852 [00:01<00:09, 78.64it/s]
 17%|█▋        | 144/852 [00:01<00:08, 78.98it/s]
 18%|█▊        | 152/852 [00:01<00:09, 77.66it/s]
 19%|█▉        | 161/852 [00:02<00:08, 79.31it/s]
 20%|█▉        | 169/852 [00:02<00:08, 79.22it/s]
 21%|██        | 178/852 [00:02<00:08, 79.74it/s]
 22%|██▏       | 187/852 [00:02<00:08, 80.55it/s]
 23%|██▎       | 196/852 [00:02<00:08, 80.55it/s]
 24%|██▍       | 205/852 [00:02<00:08, 80.53it/s]
 25%|██▌       | 214/852 [00:02<00:08, 78.05it/s]
 26%|██▌       | 223/852 [00:02<00:07, 79.41it/s]
 27%|██▋       | 232/852 [00:02<00:07, 80.51it/s]
 28%|██▊       | 241/852 [00:03<00:07, 77.55it/s]
 29%|██▉       | 250/852 [00:03<00:07, 79.18it/s]
 30%|███       | 259/852 [00:03<00:07, 80.41it/s]
 31%|███▏      | 268/852 [00:03<00:07, 80.34it/s]
 33%|███▎      | 277/852 [00:03<00:07, 81.00it/s]
 34%|███▎      | 286/852 [00:03<00:06, 81.59it/s]
 35%|███▍      | 295/852 [00:03<00:06, 80.31it/s]
 36%|███▌      | 304/852 [00:03<00:06, 81.47it/s]
 37%|███▋      | 313/852 [00:03<00:06, 79.79it/s]
 38%|███▊      | 322/852 [00:04<00:06, 81.33it/s]
 39%|███▉      | 331/852 [00:04<00:06, 80.75it/s]
 40%|███▉      | 340/852 [00:04<00:06, 80.70it/s]
 41%|████      | 349/852 [00:04<00:06, 80.37it/s]
 42%|████▏     | 358/852 [00:04<00:06, 78.68it/s]
 43%|████▎     | 367/852 [00:04<00:06, 79.85it/s]
 44%|████▍     | 376/852 [00:04<00:05, 80.53it/s]
 45%|████▌     | 385/852 [00:04<00:05, 80.46it/s]
 46%|████▌     | 394/852 [00:04<00:05, 81.26it/s]
 47%|████▋     | 403/852 [00:05<00:05, 81.11it/s]
 48%|████▊     | 412/852 [00:05<00:05, 79.03it/s]
 49%|████▉     | 421/852 [00:05<00:05, 80.48it/s]
 50%|█████     | 430/852 [00:05<00:05, 79.74it/s]
 52%|█████▏    | 439/852 [00:05<00:05, 81.17it/s]
 53%|█████▎    | 448/852 [00:05<00:04, 81.44it/s]
 54%|█████▎    | 457/852 [00:05<00:04, 82.03it/s]
 55%|█████▍    | 466/852 [00:05<00:04, 79.25it/s]
 56%|█████▌    | 474/852 [00:05<00:05, 75.40it/s]
 57%|█████▋    | 482/852 [00:06<00:04, 76.12it/s]
 58%|█████▊    | 490/852 [00:06<00:04, 76.55it/s]
 59%|█████▊    | 499/852 [00:06<00:04, 78.88it/s]
 60%|█████▉    | 507/852 [00:06<00:04, 78.87it/s]
 61%|██████    | 516/852 [00:06<00:04, 80.91it/s]
 62%|██████▏   | 525/852 [00:06<00:04, 79.57it/s]
 63%|██████▎   | 534/852 [00:06<00:03, 79.97it/s]
 64%|██████▎   | 543/852 [00:06<00:03, 81.17it/s]
 65%|██████▍   | 552/852 [00:06<00:03, 79.32it/s]
 66%|██████▌   | 561/852 [00:07<00:03, 79.73it/s]
 67%|██████▋   | 570/852 [00:07<00:03, 80.17it/s]
 68%|██████▊   | 579/852 [00:07<00:03, 79.01it/s]
 69%|██████▉   | 587/852 [00:07<00:03, 78.03it/s]
 70%|██████▉   | 596/852 [00:07<00:03, 78.78it/s]
 71%|███████   | 604/852 [00:07<00:03, 78.44it/s]
 72%|███████▏  | 612/852 [00:07<00:03, 77.03it/s]
 73%|███████▎  | 620/852 [00:07<00:02, 77.43it/s]
 74%|███████▎  | 628/852 [00:07<00:02, 77.06it/s]
 75%|███████▍  | 636/852 [00:07<00:02, 77.78it/s]
 76%|███████▌  | 644/852 [00:08<00:02, 75.78it/s]
 77%|███████▋  | 653/852 [00:08<00:02, 77.88it/s]
 78%|███████▊  | 662/852 [00:08<00:02, 78.58it/s]
 79%|███████▉  | 671/852 [00:08<00:02, 78.80it/s]
 80%|███████▉  | 680/852 [00:08<00:02, 79.13it/s]
 81%|████████  | 689/852 [00:08<00:02, 80.27it/s]
 82%|████████▏ | 698/852 [00:08<00:01, 80.46it/s]
 83%|████████▎ | 707/852 [00:08<00:01, 80.49it/s]
 84%|████████▍ | 716/852 [00:08<00:01, 80.53it/s]
 85%|████████▌ | 725/852 [00:09<00:01, 81.23it/s]
 86%|████████▌ | 734/852 [00:09<00:01, 82.34it/s]
 87%|████████▋ | 743/852 [00:09<00:01, 82.16it/s]
 88%|████████▊ | 752/852 [00:09<00:01, 82.54it/s]
 89%|████████▉ | 761/852 [00:09<00:01, 83.78it/s]
 90%|█████████ | 770/852 [00:09<00:00, 82.29it/s]
 91%|█████████▏| 779/852 [00:09<00:00, 81.45it/s]
 92%|█████████▏| 788/852 [00:09<00:00, 80.57it/s]
 94%|█████████▎| 797/852 [00:09<00:00, 80.93it/s]
 95%|█████████▍| 806/852 [00:10<00:00, 82.19it/s]
 96%|█████████▌| 815/852 [00:10<00:00, 81.02it/s]
 97%|█████████▋| 824/852 [00:10<00:00, 81.64it/s]
 98%|█████████▊| 833/852 [00:10<00:00, 82.31it/s]
 99%|█████████▉| 842/852 [00:10<00:00, 80.76it/s]
+  _warn_prf(average, modifier, msg_start, len(result))
+
+***** eval metrics *****
+  epoch                   =       10.0
+  eval_accuracy           =     0.8565
+  eval_f1                 =     0.4546
+  eval_loss               =     0.7006
+  eval_precision          =     0.3153
+  eval_recall             =     0.8145
+  eval_runtime            = 0:00:14.29
+  eval_samples            =       6810
+  eval_samples_per_second =    476.445
+  eval_steps_per_second   =     59.608
+09/06/2024 00:04:00 - INFO - __main__ -   *** Predict ***
+[INFO|trainer.py:811] 2024-09-06 00:04:00,968 >> The following columns in the test set don't have a corresponding argument in `RobertaForTokenClassification.forward` and have been ignored: id, ner_tags, tokens. If id, ner_tags, tokens are not expected by `RobertaForTokenClassification.forward`,  you can safely ignore this message.
+[INFO|trainer.py:3819] 2024-09-06 00:04:00,971 >> 
+***** Running Prediction *****
+[INFO|trainer.py:3821] 2024-09-06 00:04:00,971 >>   Num examples = 14614
+[INFO|trainer.py:3824] 2024-09-06 00:04:00,971 >>   Batch size = 8
+
  0%|          | 0/1827 [00:00<?, ?it/s]
  1%|          | 10/1827 [00:00<00:19, 91.06it/s]
  1%|          | 20/1827 [00:00<00:22, 79.50it/s]
  2%|▏         | 29/1827 [00:00<00:21, 81.87it/s]
  2%|▏         | 38/1827 [00:00<00:22, 80.70it/s]
  3%|▎         | 47/1827 [00:00<00:21, 81.06it/s]
  3%|▎         | 56/1827 [00:00<00:21, 80.92it/s]
  4%|▎         | 65/1827 [00:00<00:22, 78.97it/s]
  4%|▍         | 74/1827 [00:00<00:21, 79.80it/s]
  5%|▍         | 83/1827 [00:01<00:21, 80.85it/s]
  5%|▌         | 92/1827 [00:01<00:21, 80.71it/s]
  6%|▌         | 101/1827 [00:01<00:20, 82.27it/s]
  6%|▌         | 110/1827 [00:01<00:20, 82.50it/s]
  7%|▋         | 119/1827 [00:01<00:20, 81.39it/s]
  7%|▋         | 128/1827 [00:01<00:21, 80.71it/s]
  7%|▋         | 137/1827 [00:01<00:20, 81.92it/s]
  8%|▊         | 146/1827 [00:01<00:20, 82.51it/s]
  8%|▊         | 155/1827 [00:01<00:20, 82.36it/s]
  9%|▉         | 164/1827 [00:02<00:20, 81.09it/s]
  9%|▉         | 173/1827 [00:02<00:20, 82.25it/s]
 10%|▉         | 182/1827 [00:02<00:19, 83.50it/s]
 10%|█         | 191/1827 [00:02<00:19, 83.60it/s]
 11%|█         | 200/1827 [00:02<00:19, 82.02it/s]
 11%|█▏        | 209/1827 [00:02<00:19, 81.91it/s]
 12%|���▏        | 218/1827 [00:02<00:19, 80.50it/s]
 12%|█▏        | 227/1827 [00:02<00:19, 81.39it/s]
 13%|█▎        | 236/1827 [00:02<00:20, 78.84it/s]
 13%|█▎        | 245/1827 [00:03<00:19, 80.03it/s]
 14%|█▍        | 254/1827 [00:03<00:19, 80.43it/s]
 14%|█▍        | 263/1827 [00:03<00:19, 80.02it/s]
 15%|█▍        | 272/1827 [00:03<00:19, 81.42it/s]
 15%|█▌        | 281/1827 [00:03<00:18, 82.68it/s]
 16%|█▌        | 290/1827 [00:03<00:18, 82.21it/s]
 16%|█▋        | 299/1827 [00:03<00:18, 82.61it/s]
 17%|█▋        | 308/1827 [00:03<00:18, 82.37it/s]
 17%|█▋        | 317/1827 [00:03<00:18, 82.45it/s]
 18%|█▊        | 326/1827 [00:03<00:17, 83.97it/s]
 18%|█▊        | 335/1827 [00:04<00:17, 83.77it/s]
 19%|█▉        | 344/1827 [00:04<00:17, 84.88it/s]
 19%|█▉        | 353/1827 [00:04<00:18, 79.76it/s]
 20%|█▉        | 362/1827 [00:04<00:18, 80.32it/s]
 20%|██        | 371/1827 [00:04<00:18, 80.44it/s]
 21%|██        | 380/1827 [00:04<00:17, 80.41it/s]
 21%|██▏       | 389/1827 [00:04<00:18, 79.52it/s]
 22%|██▏       | 397/1827 [00:04<00:18, 78.70it/s]
 22%|██▏       | 406/1827 [00:04<00:17, 79.69it/s]
 23%|██▎       | 414/1827 [00:05<00:18, 77.35it/s]
 23%|██▎       | 423/1827 [00:05<00:17, 78.47it/s]
 24%|██▎       | 431/1827 [00:05<00:17, 77.94it/s]
 24%|██▍       | 439/1827 [00:05<00:17, 77.85it/s]
 24%|██▍       | 447/1827 [00:05<00:17, 77.54it/s]
 25%|██▍       | 456/1827 [00:05<00:17, 79.01it/s]
 25%|██▌       | 465/1827 [00:05<00:16, 80.47it/s]
 26%|██▌       | 474/1827 [00:05<00:17, 78.97it/s]
 26%|██▋       | 482/1827 [00:05<00:17, 77.54it/s]
 27%|██▋       | 490/1827 [00:06<00:17, 76.65it/s]
 27%|██▋       | 499/1827 [00:06<00:16, 78.31it/s]
 28%|██▊       | 508/1827 [00:06<00:16, 79.34it/s]
 28%|██▊       | 517/1827 [00:06<00:16, 79.61it/s]
 29%|██▉       | 526/1827 [00:06<00:16, 80.80it/s]
 29%|██▉       | 535/1827 [00:06<00:16, 80.00it/s]
 30%|██▉       | 544/1827 [00:06<00:15, 80.92it/s]
 30%|███       | 553/1827 [00:06<00:15, 80.34it/s]
 31%|███       | 562/1827 [00:06<00:15, 81.20it/s]
 31%|███▏      | 571/1827 [00:07<00:15, 83.07it/s]
 32%|███▏      | 580/1827 [00:07<00:14, 83.38it/s]
 32%|███▏      | 589/1827 [00:07<00:14, 83.03it/s]
 33%|███▎      | 598/1827 [00:07<00:14, 82.25it/s]
 33%|███▎      | 607/1827 [00:07<00:14, 81.98it/s]
 34%|███▎      | 616/1827 [00:07<00:14, 81.52it/s]
 34%|███▍      | 625/1827 [00:07<00:14, 82.30it/s]
 35%|███▍      | 634/1827 [00:07<00:14, 82.87it/s]
 35%|███▌      | 643/1827 [00:07<00:14, 83.83it/s]
 36%|███▌      | 652/1827 [00:08<00:15, 78.28it/s]
 36%|███▌      | 661/1827 [00:08<00:14, 79.08it/s]
 37%|███▋      | 670/1827 [00:08<00:14, 80.28it/s]
 37%|███▋      | 679/1827 [00:08<00:14, 80.78it/s]
 38%|███▊      | 688/1827 [00:08<00:13, 82.11it/s]
 38%|███▊      | 697/1827 [00:08<00:13, 83.76it/s]
 39%|███▊      | 706/1827 [00:08<00:13, 83.15it/s]
 39%|███▉      | 715/1827 [00:08<00:13, 84.66it/s]
 40%|███▉      | 724/1827 [00:08<00:12, 85.34it/s]
 40%|████      | 733/1827 [00:09<00:12, 85.46it/s]
 41%|████      | 742/1827 [00:09<00:12, 85.34it/s]
 41%|████      | 751/1827 [00:09<00:12, 86.45it/s]
 42%|████▏     | 760/1827 [00:09<00:12, 85.05it/s]
 42%|████▏     | 769/1827 [00:09<00:12, 86.03it/s]
 43%|████▎     | 778/1827 [00:09<00:12, 84.73it/s]
 43%|████▎     | 787/1827 [00:09<00:12, 81.97it/s]
 44%|████▎     | 796/1827 [00:09<00:12, 83.05it/s]
 44%|████▍     | 805/1827 [00:09<00:12, 81.61it/s]
 45%|████▍     | 814/1827 [00:09<00:12, 82.53it/s]
 45%|████▌     | 823/1827 [00:10<00:12, 82.99it/s]
 46%|████▌     | 832/1827 [00:10<00:12, 82.38it/s]
 46%|████▌     | 841/1827 [00:10<00:11, 82.96it/s]
 47%|████▋     | 850/1827 [00:10<00:11, 83.72it/s]
 47%|████▋     | 859/1827 [00:10<00:11, 84.83it/s]
 48%|████▊     | 868/1827 [00:10<00:11, 84.11it/s]
 48%|████▊     | 877/1827 [00:10<00:11, 83.69it/s]
 48%|████▊     | 886/1827 [00:10<00:11, 84.15it/s]
 49%|████▉     | 895/1827 [00:10<00:11, 83.37it/s]
 49%|████▉     | 904/1827 [00:11<00:10, 84.07it/s]
 50%|████▉     | 913/1827 [00:11<00:10, 84.82it/s]
 50%|█████     | 922/1827 [00:11<00:10, 84.51it/s]
 51%|█████     | 931/1827 [00:11<00:10, 81.89it/s]
 51%|█████▏    | 940/1827 [00:11<00:10, 82.43it/s]
 52%|█████▏    | 949/1827 [00:11<00:10, 81.02it/s]
 52%|█████▏    | 958/1827 [00:11<00:10, 81.56it/s]
 53%|█████▎    | 967/1827 [00:11<00:10, 83.24it/s]
 53%|█████▎    | 976/1827 [00:11<00:10, 82.41it/s]
 54%|█████▍    | 985/1827 [00:12<00:10, 83.96it/s]
 54%|█████▍    | 994/1827 [00:12<00:09, 84.30it/s]
 55%|█████▍    | 1003/1827 [00:12<00:09, 83.87it/s]
 55%|█████▌    | 1012/1827 [00:12<00:09, 83.40it/s]
 56%|█████▌    | 1021/1827 [00:12<00:09, 84.27it/s]
 56%|█████▋    | 1030/1827 [00:12<00:09, 84.53it/s]
 57%|█████▋    | 1039/1827 [00:12<00:09, 82.78it/s]
 57%|█████▋    | 1048/1827 [00:12<00:09, 83.18it/s]
 58%|█████▊    | 1057/1827 [00:12<00:09, 83.19it/s]
 58%|█████▊    | 1066/1827 [00:13<00:09, 84.07it/s]
 59%|█████▉    | 1075/1827 [00:13<00:08, 84.64it/s]
 59%|█████▉    | 1084/1827 [00:13<00:08, 85.01it/s]
 60%|█████▉    | 1093/1827 [00:13<00:08, 85.36it/s]
 60%|██████    | 1102/1827 [00:13<00:08, 84.17it/s]
 61%|██████    | 1111/1827 [00:13<00:08, 82.68it/s]
 61%|██████▏   | 1120/1827 [00:13<00:08, 81.76it/s]
 62%|██████▏   | 1129/1827 [00:13<00:08, 81.92it/s]
 62%|██████▏   | 1138/1827 [00:13<00:08, 81.83it/s]
 63%|██████▎   | 1147/1827 [00:13<00:08, 82.31it/s]
 63%|██████▎   | 1156/1827 [00:14<00:08, 82.02it/s]
 64%|██████▍   | 1165/1827 [00:14<00:08, 79.48it/s]
 64%|██████▍   | 1174/1827 [00:14<00:08, 80.60it/s]
 65%|██████▍   | 1183/1827 [00:14<00:08, 78.91it/s]
 65%|██████▌   | 1192/1827 [00:14<00:07, 81.02it/s]
 66%|██████▌   | 1201/1827 [00:14<00:07, 80.79it/s]
 66%|██████▌   | 1210/1827 [00:14<00:07, 81.06it/s]
 67%|██████▋   | 1219/1827 [00:14<00:07, 79.89it/s]
 67%|██████▋   | 1228/1827 [00:14<00:07, 81.29it/s]
 68%|██████▊   | 1237/1827 [00:15<00:07, 82.50it/s]
 68%|██████▊   | 1246/1827 [00:15<00:07, 82.71it/s]
 69%|██████▊   | 1255/1827 [00:15<00:07, 80.01it/s]
 69%|██████▉   | 1264/1827 [00:15<00:07, 80.07it/s]
 70%|██████▉   | 1273/1827 [00:15<00:06, 81.19it/s]
 70%|███████   | 1282/1827 [00:15<00:06, 82.47it/s]
 71%|███████   | 1291/1827 [00:15<00:06, 83.35it/s]
 71%|███████   | 1300/1827 [00:15<00:06, 83.34it/s]
 72%|███████▏  | 1309/1827 [00:15<00:06, 83.88it/s]
 72%|███████▏  | 1318/1827 [00:16<00:05, 84.84it/s]
 73%|███████▎  | 1327/1827 [00:16<00:05, 83.90it/s]
 73%|███████▎  | 1336/1827 [00:16<00:05, 84.12it/s]
 74%|███████▎  | 1345/1827 [00:16<00:05, 83.22it/s]
 74%|███████▍  | 1354/1827 [00:16<00:05, 82.63it/s]
 75%|███████▍  | 1363/1827 [00:16<00:05, 82.32it/s]
 75%|███████▌  | 1372/1827 [00:16<00:05, 81.87it/s]
 76%|███████▌  | 1381/1827 [00:16<00:05, 82.35it/s]
 76%|███████▌  | 1390/1827 [00:16<00:05, 81.43it/s]
 77%|███████▋  | 1399/1827 [00:17<00:05, 82.08it/s]
 77%|███████▋  | 1408/1827 [00:17<00:05, 82.67it/s]
 78%|███████▊  | 1417/1827 [00:17<00:04, 82.84it/s]
 78%|███████▊  | 1426/1827 [00:17<00:04, 82.22it/s]
 79%|███████▊  | 1435/1827 [00:17<00:04, 81.18it/s]
 79%|███████▉  | 1444/1827 [00:17<00:04, 78.91it/s]
 80%|███████▉  | 1453/1827 [00:17<00:04, 80.47it/s]
 80%|████████  | 1462/1827 [00:17<00:04, 79.66it/s]
 80%|████████  | 1470/1827 [00:17<00:04, 78.26it/s]
 81%|████████  | 1478/1827 [00:18<00:04, 78.65it/s]
 81%|████████▏ | 1487/1827 [00:18<00:04, 79.78it/s]
 82%|████████▏ | 1495/1827 [00:18<00:04, 77.95it/s]
 82%|████████▏ | 1503/1827 [00:18<00:04, 77.75it/s]
 83%|████████▎ | 1512/1827 [00:18<00:03, 79.71it/s]
 83%|████████▎ | 1521/1827 [00:18<00:03, 80.57it/s]
 84%|████████▎ | 1530/1827 [00:18<00:03, 80.96it/s]
 84%|████████▍ | 1539/1827 [00:18<00:03, 81.58it/s]
 85%|████████▍ | 1548/1827 [00:18<00:03, 81.13it/s]
 85%|████████▌ | 1557/1827 [00:19<00:03, 81.30it/s]
 86%|████████▌ | 1566/1827 [00:19<00:03, 81.60it/s]
 86%|████████▌ | 1575/1827 [00:19<00:03, 81.67it/s]
 87%|████████▋ | 1584/1827 [00:19<00:02, 82.53it/s]
 87%|████████▋ | 1593/1827 [00:19<00:02, 82.02it/s]
 88%|████████▊ | 1602/1827 [00:19<00:02, 82.22it/s]
 88%|████████▊ | 1611/1827 [00:19<00:02, 82.92it/s]
 89%|████████▊ | 1620/1827 [00:19<00:02, 79.31it/s]
 89%|████████▉ | 1628/1827 [00:19<00:02, 76.38it/s]
 90%|████████▉ | 1637/1827 [00:20<00:02, 78.37it/s]
 90%|█████████ | 1646/1827 [00:20<00:02, 79.98it/s]
 91%|█████████ | 1655/1827 [00:20<00:02, 81.60it/s]
 91%|█████████ | 1664/1827 [00:20<00:02, 81.06it/s]
 92%|█████████▏| 1673/1827 [00:20<00:01, 81.37it/s]
 92%|█████████▏| 1682/1827 [00:20<00:01, 82.25it/s]
 93%|█████████▎| 1691/1827 [00:20<00:01, 83.16it/s]
 93%|█████████▎| 1700/1827 [00:20<00:01, 81.57it/s]
 94%|█████████▎| 1709/1827 [00:20<00:01, 82.04it/s]
 94%|█████████▍| 1718/1827 [00:21<00:01, 83.23it/s]
 95%|█████████▍| 1727/1827 [00:21<00:01, 80.36it/s]
 95%|█████████▌| 1736/1827 [00:21<00:01, 81.04it/s]
 96%|█████████▌| 1745/1827 [00:21<00:01, 81.89it/s]
 96%|█████████▌| 1754/1827 [00:21<00:00, 82.74it/s]
 96%|█████████▋| 1763/1827 [00:21<00:00, 82.25it/s]
 97%|█████████▋| 1772/1827 [00:21<00:00, 82.64it/s]
 97%|█████████▋| 1781/1827 [00:21<00:00, 83.22it/s]
 98%|█████████▊| 1790/1827 [00:21<00:00, 83.23it/s]
 98%|█████████▊| 1799/1827 [00:22<00:00, 80.14it/s]
 99%|█████████▉| 1808/1827 [00:22<00:00, 81.33it/s]
 99%|█████████▉| 1817/1827 [00:22<00:00, 81.79it/s]
+[INFO|trainer.py:3503] 2024-09-06 00:04:30,726 >> Saving model checkpoint to /content/dissertation/scripts/ner/output
+[INFO|configuration_utils.py:472] 2024-09-06 00:04:30,727 >> Configuration saved in /content/dissertation/scripts/ner/output/config.json
+[INFO|modeling_utils.py:2799] 2024-09-06 00:04:32,122 >> Model weights saved in /content/dissertation/scripts/ner/output/model.safetensors
+[INFO|tokenization_utils_base.py:2684] 2024-09-06 00:04:32,123 >> tokenizer config file saved in /content/dissertation/scripts/ner/output/tokenizer_config.json
+[INFO|tokenization_utils_base.py:2693] 2024-09-06 00:04:32,123 >> Special tokens file saved in /content/dissertation/scripts/ner/output/special_tokens_map.json
+***** predict metrics *****
+  predict_accuracy           =     0.9438
+  predict_f1                 =     0.6366
+  predict_loss               =     0.2277
+  predict_precision          =     0.5196
+  predict_recall             =     0.8216
+  predict_runtime            = 0:00:29.10
+  predict_samples_per_second =    502.103
+  predict_steps_per_second   =     62.771
+

train_results.json

@@ -1,9 +1,9 @@
 {
     "epoch": 10.0,
-    "total_flos": 1.4714840952259542e+16,
-    "train_loss": 0.002772659832779558,
-    "train_runtime": 1349.0548,
-    "train_samples": 29797,
-    "train_samples_per_second": 220.873,
-    "train_steps_per_second": 3.454
+    "total_flos": 1.7176580067661056e+16,
+    "train_loss": 0.0812657987344287,
+    "train_runtime": 1549.44,
+    "train_samples": 34604,
+    "train_samples_per_second": 223.332,
+    "train_steps_per_second": 3.492
 }

trainer_state.json

@@ -1,208 +1,215 @@
 {
-  "best_metric": 0.9491525423728814,
-  "best_model_checkpoint": "/content/dissertation/scripts/ner/output/checkpoint-2330",
+  "best_metric": 0.45455732567249935,
+  "best_model_checkpoint": "/content/dissertation/scripts/ner/output/checkpoint-2705",
   "epoch": 10.0,
   "eval_steps": 500,
-  "global_step": 4660,
+  "global_step": 5410,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 1.0,
-      "eval_accuracy": 0.9989495654084337,
-      "eval_f1": 0.9351598173515981,
-      "eval_loss": 0.0030670168343931437,
-      "eval_precision": 0.9292196007259528,
-      "eval_recall": 0.9411764705882353,
-      "eval_runtime": 13.9946,
-      "eval_samples_per_second": 486.615,
-      "eval_steps_per_second": 60.88,
-      "step": 466
-    },
-    {
-      "epoch": 1.0729613733905579,
-      "grad_norm": 0.07174628973007202,
-      "learning_rate": 4.4635193133047216e-05,
-      "loss": 0.0199,
+      "epoch": 0.9242144177449169,
+      "grad_norm": 1.4528056383132935,
+      "learning_rate": 4.537892791127542e-05,
+      "loss": 0.3191,
       "step": 500
     },
     {
-      "epoch": 2.0,
-      "eval_accuracy": 0.9989358342373021,
-      "eval_f1": 0.9386834986474301,
-      "eval_loss": 0.0030621723271906376,
-      "eval_precision": 0.9212389380530973,
-      "eval_recall": 0.9568014705882353,
-      "eval_runtime": 14.0045,
-      "eval_samples_per_second": 486.272,
-      "eval_steps_per_second": 60.838,
-      "step": 932
-    },
-    {
-      "epoch": 2.1459227467811157,
-      "grad_norm": 0.13906870782375336,
-      "learning_rate": 3.927038626609442e-05,
-      "loss": 0.0026,
+      "epoch": 1.0,
+      "eval_accuracy": 0.8443022505389485,
+      "eval_f1": 0.407486125870823,
+      "eval_loss": 0.47723615169525146,
+      "eval_precision": 0.27250473783954515,
+      "eval_recall": 0.8074403369209172,
+      "eval_runtime": 14.3908,
+      "eval_samples_per_second": 473.218,
+      "eval_steps_per_second": 59.204,
+      "step": 541
+    },
+    {
+      "epoch": 1.8484288354898335,
+      "grad_norm": 1.5587466955184937,
+      "learning_rate": 4.075785582255083e-05,
+      "loss": 0.1619,
       "step": 1000
     },
     {
-      "epoch": 3.0,
-      "eval_accuracy": 0.9989152374806047,
-      "eval_f1": 0.9360919540229885,
-      "eval_loss": 0.004003152716904879,
-      "eval_precision": 0.9365225390984361,
-      "eval_recall": 0.9356617647058824,
-      "eval_runtime": 13.9451,
-      "eval_samples_per_second": 488.343,
-      "eval_steps_per_second": 61.097,
-      "step": 1398
-    },
-    {
-      "epoch": 3.218884120171674,
-      "grad_norm": 0.14111244678497314,
-      "learning_rate": 3.3905579399141636e-05,
-      "loss": 0.0011,
+      "epoch": 2.0,
+      "eval_accuracy": 0.8552871874442172,
+      "eval_f1": 0.43975123088883133,
+      "eval_loss": 0.4583655595779419,
+      "eval_precision": 0.30406737143881024,
+      "eval_recall": 0.7941038839494619,
+      "eval_runtime": 14.212,
+      "eval_samples_per_second": 479.173,
+      "eval_steps_per_second": 59.949,
+      "step": 1082
+    },
+    {
+      "epoch": 2.7726432532347505,
+      "grad_norm": 0.9616082310676575,
+      "learning_rate": 3.613678373382625e-05,
+      "loss": 0.11,
       "step": 1500
     },
     {
-      "epoch": 4.0,
-      "eval_accuracy": 0.9987230010847625,
-      "eval_f1": 0.9308584686774942,
-      "eval_loss": 0.005216046702116728,
-      "eval_precision": 0.9400187441424555,
-      "eval_recall": 0.921875,
-      "eval_runtime": 14.2384,
-      "eval_samples_per_second": 478.283,
-      "eval_steps_per_second": 59.838,
-      "step": 1864
-    },
-    {
-      "epoch": 4.291845493562231,
-      "grad_norm": 0.09000600874423981,
-      "learning_rate": 2.8540772532188842e-05,
-      "loss": 0.001,
+      "epoch": 3.0,
+      "eval_accuracy": 0.8435470361267112,
+      "eval_f1": 0.4338006724608259,
+      "eval_loss": 0.6447410583496094,
+      "eval_precision": 0.29758899817216466,
+      "eval_recall": 0.7999532054281703,
+      "eval_runtime": 14.3233,
+      "eval_samples_per_second": 475.451,
+      "eval_steps_per_second": 59.484,
+      "step": 1623
+    },
+    {
+      "epoch": 3.6968576709796674,
+      "grad_norm": 1.4018645286560059,
+      "learning_rate": 3.1515711645101665e-05,
+      "loss": 0.0764,
       "step": 2000
     },
     {
-      "epoch": 5.0,
-      "eval_accuracy": 0.9989426998228679,
-      "eval_f1": 0.9491525423728814,
-      "eval_loss": 0.004777050111442804,
-      "eval_precision": 0.9461187214611873,
-      "eval_recall": 0.9522058823529411,
-      "eval_runtime": 13.9397,
-      "eval_samples_per_second": 488.533,
-      "eval_steps_per_second": 61.12,
-      "step": 2330
-    },
-    {
-      "epoch": 5.364806866952789,
-      "grad_norm": 0.0045097870752215385,
-      "learning_rate": 2.3175965665236052e-05,
-      "loss": 0.0005,
+      "epoch": 4.0,
+      "eval_accuracy": 0.8398602166778805,
+      "eval_f1": 0.42338430558177587,
+      "eval_loss": 0.7413247227668762,
+      "eval_precision": 0.2895756219333735,
+      "eval_recall": 0.7870846981750117,
+      "eval_runtime": 14.5396,
+      "eval_samples_per_second": 468.375,
+      "eval_steps_per_second": 58.598,
+      "step": 2164
+    },
+    {
+      "epoch": 4.621072088724584,
+      "grad_norm": 1.0904265642166138,
+      "learning_rate": 2.6894639556377083e-05,
+      "loss": 0.0567,
       "step": 2500
     },
     {
-      "epoch": 6.0,
-      "eval_accuracy": 0.9988809095527758,
-      "eval_f1": 0.9448244414044688,
-      "eval_loss": 0.004644877277314663,
-      "eval_precision": 0.9375565610859729,
-      "eval_recall": 0.9522058823529411,
-      "eval_runtime": 14.0259,
-      "eval_samples_per_second": 485.531,
-      "eval_steps_per_second": 60.745,
-      "step": 2796
-    },
-    {
-      "epoch": 6.437768240343348,
-      "grad_norm": 0.029634617269039154,
-      "learning_rate": 1.7811158798283262e-05,
-      "loss": 0.0004,
+      "epoch": 5.0,
+      "eval_accuracy": 0.8564886649182308,
+      "eval_f1": 0.45455732567249935,
+      "eval_loss": 0.7005925178527832,
+      "eval_precision": 0.3152508603513856,
+      "eval_recall": 0.8144595226953674,
+      "eval_runtime": 14.2723,
+      "eval_samples_per_second": 477.148,
+      "eval_steps_per_second": 59.696,
+      "step": 2705
+    },
+    {
+      "epoch": 5.545286506469501,
+      "grad_norm": 1.240962028503418,
+      "learning_rate": 2.2273567467652497e-05,
+      "loss": 0.0428,
       "step": 3000
     },
     {
-      "epoch": 7.0,
-      "eval_accuracy": 0.9989564309939994,
-      "eval_f1": 0.9446460980036298,
-      "eval_loss": 0.0049773636274039745,
-      "eval_precision": 0.9327956989247311,
-      "eval_recall": 0.9568014705882353,
-      "eval_runtime": 13.9217,
-      "eval_samples_per_second": 489.166,
-      "eval_steps_per_second": 61.2,
-      "step": 3262
-    },
-    {
-      "epoch": 7.510729613733906,
-      "grad_norm": 0.003798937890678644,
-      "learning_rate": 1.2446351931330473e-05,
-      "loss": 0.0002,
+      "epoch": 6.0,
+      "eval_accuracy": 0.8504057561069384,
+      "eval_f1": 0.44700636942675154,
+      "eval_loss": 0.8111857175827026,
+      "eval_precision": 0.30710659898477155,
+      "eval_recall": 0.8210107627515209,
+      "eval_runtime": 14.3991,
+      "eval_samples_per_second": 472.946,
+      "eval_steps_per_second": 59.17,
+      "step": 3246
+    },
+    {
+      "epoch": 6.469500924214418,
+      "grad_norm": 0.44737720489501953,
+      "learning_rate": 1.7652495378927914e-05,
+      "loss": 0.0332,
       "step": 3500
     },
     {
-      "epoch": 8.0,
-      "eval_accuracy": 0.9989221030661705,
-      "eval_f1": 0.9435520881138136,
-      "eval_loss": 0.005484889727085829,
-      "eval_precision": 0.9422548120989918,
-      "eval_recall": 0.9448529411764706,
-      "eval_runtime": 13.8923,
-      "eval_samples_per_second": 490.2,
-      "eval_steps_per_second": 61.329,
-      "step": 3728
-    },
-    {
-      "epoch": 8.583690987124463,
-      "grad_norm": 0.0004711664514616132,
-      "learning_rate": 7.0815450643776825e-06,
-      "loss": 0.0001,
+      "epoch": 7.0,
+      "eval_accuracy": 0.8532961676301372,
+      "eval_f1": 0.4493518337567586,
+      "eval_loss": 0.904643714427948,
+      "eval_precision": 0.3113658932924077,
+      "eval_recall": 0.8069723912026205,
+      "eval_runtime": 14.3036,
+      "eval_samples_per_second": 476.105,
+      "eval_steps_per_second": 59.566,
+      "step": 3787
+    },
+    {
+      "epoch": 7.393715341959335,
+      "grad_norm": 0.7116318941116333,
+      "learning_rate": 1.3031423290203328e-05,
+      "loss": 0.0257,
       "step": 4000
     },
     {
-      "epoch": 9.0,
-      "eval_accuracy": 0.9989426998228679,
-      "eval_f1": 0.9441903019213176,
-      "eval_loss": 0.0057435426861047745,
-      "eval_precision": 0.9398907103825137,
-      "eval_recall": 0.9485294117647058,
-      "eval_runtime": 13.9755,
-      "eval_samples_per_second": 487.281,
-      "eval_steps_per_second": 60.964,
-      "step": 4194
-    },
-    {
-      "epoch": 9.656652360515022,
-      "grad_norm": 0.005946693476289511,
-      "learning_rate": 1.7167381974248929e-06,
-      "loss": 0.0001,
+      "epoch": 8.0,
+      "eval_accuracy": 0.8481538440413583,
+      "eval_f1": 0.44435897435897437,
+      "eval_loss": 0.9722912907600403,
+      "eval_precision": 0.30602154335158044,
+      "eval_recall": 0.8109499298081423,
+      "eval_runtime": 14.4222,
+      "eval_samples_per_second": 472.19,
+      "eval_steps_per_second": 59.076,
+      "step": 4328
+    },
+    {
+      "epoch": 8.317929759704251,
+      "grad_norm": 0.9520462155342102,
+      "learning_rate": 8.410351201478742e-06,
+      "loss": 0.022,
       "step": 4500
     },
+    {
+      "epoch": 9.0,
+      "eval_accuracy": 0.850186057368833,
+      "eval_f1": 0.44668737060041414,
+      "eval_loss": 1.002764105796814,
+      "eval_precision": 0.30871042747272404,
+      "eval_recall": 0.8076743097800655,
+      "eval_runtime": 14.2485,
+      "eval_samples_per_second": 477.944,
+      "eval_steps_per_second": 59.796,
+      "step": 4869
+    },
+    {
+      "epoch": 9.242144177449168,
+      "grad_norm": 0.6707109212875366,
+      "learning_rate": 3.789279112754159e-06,
+      "loss": 0.0181,
+      "step": 5000
+    },
     {
       "epoch": 10.0,
-      "eval_accuracy": 0.9989083718950389,
-      "eval_f1": 0.9416058394160585,
-      "eval_loss": 0.005793666001409292,
-      "eval_precision": 0.9347826086956522,
-      "eval_recall": 0.9485294117647058,
-      "eval_runtime": 14.4595,
-      "eval_samples_per_second": 470.971,
-      "eval_steps_per_second": 58.923,
-      "step": 4660
+      "eval_accuracy": 0.8533304955579661,
+      "eval_f1": 0.45038613797131544,
+      "eval_loss": 1.0022608041763306,
+      "eval_precision": 0.31162999550965426,
+      "eval_recall": 0.8118858212447356,
+      "eval_runtime": 14.4559,
+      "eval_samples_per_second": 471.089,
+      "eval_steps_per_second": 58.938,
+      "step": 5410
     },
     {
       "epoch": 10.0,
-      "step": 4660,
-      "total_flos": 1.4714840952259542e+16,
-      "train_loss": 0.002772659832779558,
-      "train_runtime": 1349.0548,
-      "train_samples_per_second": 220.873,
-      "train_steps_per_second": 3.454
+      "step": 5410,
+      "total_flos": 1.7176580067661056e+16,
+      "train_loss": 0.0812657987344287,
+      "train_runtime": 1549.44,
+      "train_samples_per_second": 223.332,
+      "train_steps_per_second": 3.492
     }
   ],
   "logging_steps": 500,
-  "max_steps": 4660,
+  "max_steps": 5410,
   "num_input_tokens_seen": 0,
   "num_train_epochs": 10,
   "save_steps": 500,
@@ -218,7 +225,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 1.4714840952259542e+16,
+  "total_flos": 1.7176580067661056e+16,
   "train_batch_size": 32,
   "trial_name": null,
   "trial_params": null