End of training

README.md

@@ -25,16 +25,16 @@ model-index:
     metrics:
     - name: Precision
       type: precision
-      value: 0.937247539398077
+      value: 0.9497881598534296
     - name: Recall
       type: recall
-      value: 0.964689348246179
+      value: 0.9714235521461615
     - name: F1
       type: f1
-      value: 0.9507704738269752
+      value: 0.9604840343919173
     - name: Accuracy
       type: accuracy
-      value: 0.9773235561218265
+      value: 0.981362755330252
 ---
 
 <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -44,11 +44,11 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [raulgdp/xml-roberta-large-finetuned-ner](https://huggingface.co/raulgdp/xml-roberta-large-finetuned-ner) on the biobert_json dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.1065
-- Precision: 0.9372
-- Recall: 0.9647
-- F1: 0.9508
-- Accuracy: 0.9773
+- Loss: 0.0946
+- Precision: 0.9498
+- Recall: 0.9714
+- F1: 0.9605
+- Accuracy: 0.9814
 
 ## Model description
 
@@ -68,8 +68,8 @@ More information needed
 
 The following hyperparameters were used during training:
 - learning_rate: 2e-05
-- train_batch_size: 4
-- eval_batch_size: 4
+- train_batch_size: 8
+- eval_batch_size: 8
 - seed: 42
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: linear
@@ -77,13 +77,13 @@ The following hyperparameters were used during training:
 
 ### Training results
 
-| Training Loss | Epoch | Step  | Validation Loss | Precision | Recall | F1     | Accuracy |
-|:-------------:|:-----:|:-----:|:---------------:|:---------:|:------:|:------:|:--------:|
-| 0.1305        | 1.0   | 2447  | 0.1005          | 0.9298    | 0.9680 | 0.9485 | 0.9747   |
-| 0.0874        | 2.0   | 4894  | 0.0981          | 0.9406    | 0.9711 | 0.9556 | 0.9781   |
-| 0.0782        | 3.0   | 7341  | 0.1023          | 0.9245    | 0.9577 | 0.9408 | 0.9747   |
-| 0.0807        | 4.0   | 9788  | 0.1042          | 0.9316    | 0.9567 | 0.9440 | 0.9753   |
-| 0.0437        | 5.0   | 12235 | 0.1065          | 0.9372    | 0.9647 | 0.9508 | 0.9773   |
+| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1     | Accuracy |
+|:-------------:|:-----:|:----:|:---------------:|:---------:|:------:|:------:|:--------:|
+| 0.1306        | 1.0   | 1224 | 0.1013          | 0.9299    | 0.9609 | 0.9451 | 0.9735   |
+| 0.0996        | 2.0   | 2448 | 0.0932          | 0.9383    | 0.9656 | 0.9517 | 0.9777   |
+| 0.0608        | 3.0   | 3672 | 0.0865          | 0.9493    | 0.9720 | 0.9605 | 0.9813   |
+| 0.0445        | 4.0   | 4896 | 0.0927          | 0.9531    | 0.9729 | 0.9629 | 0.9819   |
+| 0.0327        | 5.0   | 6120 | 0.0946          | 0.9498    | 0.9714 | 0.9605 | 0.9814   |
 
 
 ### Framework versions