pabRomero
/

BERT-full-finetuned-ner-pablo

Token Classification

Generated from Trainer

Inference Endpoints

Model card Files Files and versions Metrics Training metrics Community

pabRomero commited on Aug 16, 2024

Commit

a38233f

·

verified ·

1 Parent(s): ddad95f

Training complete

Files changed (1) hide show

README.md +14 -14

README.md CHANGED Viewed

@@ -19,11 +19,11 @@ should probably proofread and complete it, then remove this comment. -->
 This model is a fine-tuned version of [medicalai/ClinicalBERT](https://huggingface.co/medicalai/ClinicalBERT) on the None dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.1039
-- Precision: 0.7571
-- Recall: 0.7522
-- F1: 0.7546
-- Accuracy: 0.9725
 ## Model description
@@ -43,24 +43,24 @@ More information needed
 The following hyperparameters were used during training:
 - learning_rate: 0.0002
-- train_batch_size: 8
-- eval_batch_size: 8
 - seed: 42
-- gradient_accumulation_steps: 2
 - total_train_batch_size: 16
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: linear
-- lr_scheduler_warmup_ratio: 0.1
 - num_epochs: 3
 - mixed_precision_training: Native AMP
 ### Training results
-| Training Loss | Epoch  | Step  | Validation Loss | Precision | Recall | F1     | Accuracy |
-|:-------------:|:------:|:-----:|:---------------:|:---------:|:------:|:------:|:--------:|
-| 0.1134        | 0.9999 | 4956  | 0.1136          | 0.7459    | 0.7336 | 0.7397 | 0.9697   |
-| 0.1001        | 2.0    | 9913  | 0.1060          | 0.7442    | 0.7533 | 0.7487 | 0.9717   |
-| 0.1024        | 2.9997 | 14868 | 0.1039          | 0.7571    | 0.7522 | 0.7546 | 0.9725   |
 ### Framework versions

 This model is a fine-tuned version of [medicalai/ClinicalBERT](https://huggingface.co/medicalai/ClinicalBERT) on the None dataset.
 It achieves the following results on the evaluation set:
+- Loss: 0.1117
+- Precision: 0.8051
+- Recall: 0.7944
+- F1: 0.7997
+- Accuracy: 0.9702
 ## Model description
 The following hyperparameters were used during training:
 - learning_rate: 0.0002
+- train_batch_size: 4
+- eval_batch_size: 4
 - seed: 42
+- gradient_accumulation_steps: 4
 - total_train_batch_size: 16
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: linear
+- lr_scheduler_warmup_ratio: 0.05
 - num_epochs: 3
 - mixed_precision_training: Native AMP
 ### Training results
+| Training Loss | Epoch  | Step | Validation Loss | Precision | Recall | F1     | Accuracy |
+|:-------------:|:------:|:----:|:---------------:|:---------:|:------:|:------:|:--------:|
+| 0.135         | 0.9998 | 2351 | 0.1292          | 0.7596    | 0.7329 | 0.7460 | 0.9649   |
+| 0.0863        | 2.0    | 4703 | 0.1222          | 0.8064    | 0.7631 | 0.7841 | 0.9690   |
+| 0.0554        | 2.9994 | 7053 | 0.1117          | 0.8051    | 0.7944 | 0.7997 | 0.9702   |
 ### Framework versions