Model save

README.md

@@ -18,8 +18,8 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [distilbert/distilbert-base-uncased](https://huggingface.co/distilbert/distilbert-base-uncased) on an unknown dataset.
 It achieves the following results on the evaluation set:
-- Loss: 1.3081
-- Accuracy: 0.7230
+- Loss: 0.6824
+- Accuracy: 0.7670
 
 ## Model description
 
@@ -38,52 +38,26 @@ More information needed
 ### Training hyperparameters
 
 The following hyperparameters were used during training:
-- learning_rate: 1e-05
+- learning_rate: 2e-06
 - train_batch_size: 32
 - eval_batch_size: 8
 - seed: 42
 - optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
 - lr_scheduler_type: linear
-- lr_scheduler_warmup_steps: 500
-- num_epochs: 32
+- lr_scheduler_warmup_ratio: 0.1
+- num_epochs: 6
 - mixed_precision_training: Native AMP
 
 ### Training results
 
-| Training Loss | Epoch | Step  | Validation Loss | Accuracy |
-|:-------------:|:-----:|:-----:|:---------------:|:--------:|
-| 1.5456        | 1.0   | 876   | 0.6853          | 0.7600   |
-| 0.6286        | 2.0   | 1752  | 0.6386          | 0.7755   |
-| 0.5239        | 3.0   | 2628  | 0.6216          | 0.7690   |
-| 0.4708        | 4.0   | 3504  | 0.6294          | 0.7684   |
-| 0.4224        | 5.0   | 4380  | 0.6622          | 0.7690   |
-| 0.3963        | 6.0   | 5256  | 0.6938          | 0.7558   |
-| 0.3677        | 7.0   | 6132  | 0.7313          | 0.7575   |
-| 0.3486        | 8.0   | 7008  | 0.7668          | 0.7508   |
-| 0.3192        | 9.0   | 7884  | 0.7856          | 0.7488   |
-| 0.3119        | 10.0  | 8760  | 0.8193          | 0.7413   |
-| 0.297         | 11.0  | 9636  | 0.8250          | 0.7385   |
-| 0.2826        | 12.0  | 10512 | 0.8906          | 0.7264   |
-| 0.2664        | 13.0  | 11388 | 0.8942          | 0.7301   |
-| 0.2614        | 14.0  | 12264 | 0.9402          | 0.7281   |
-| 0.2585        | 15.0  | 13140 | 0.9722          | 0.7361   |
-| 0.2603        | 16.0  | 14016 | 1.0199          | 0.7285   |
-| 0.2366        | 17.0  | 14892 | 1.0044          | 0.7290   |
-| 0.2406        | 18.0  | 15768 | 1.0022          | 0.7237   |
-| 0.2341        | 19.0  | 16644 | 1.0498          | 0.7233   |
-| 0.2392        | 20.0  | 17520 | 1.0741          | 0.7258   |
-| 0.2135        | 21.0  | 18396 | 1.1113          | 0.7233   |
-| 0.2166        | 22.0  | 19272 | 1.1229          | 0.7288   |
-| 0.216         | 23.0  | 20148 | 1.1429          | 0.7204   |
-| 0.2164        | 24.0  | 21024 | 1.1872          | 0.7261   |
-| 0.2079        | 25.0  | 21900 | 1.2140          | 0.7161   |
-| 0.2027        | 26.0  | 22776 | 1.2285          | 0.7240   |
-| 0.201         | 27.0  | 23652 | 1.2247          | 0.7250   |
-| 0.2052        | 28.0  | 24528 | 1.2356          | 0.7207   |
-| 0.1959        | 29.0  | 25404 | 1.2895          | 0.7217   |
-| 0.1975        | 30.0  | 26280 | 1.2861          | 0.7245   |
-| 0.1937        | 31.0  | 27156 | 1.3034          | 0.7211   |
-| 0.194         | 32.0  | 28032 | 1.3081          | 0.7230   |
+| Training Loss | Epoch | Step | Validation Loss | Accuracy |
+|:-------------:|:-----:|:----:|:---------------:|:--------:|
+| 1.9725        | 1.0   | 876  | 1.0953          | 0.6834   |
+| 0.9389        | 2.0   | 1752 | 0.7982          | 0.7514   |
+| 0.7549        | 3.0   | 2628 | 0.7214          | 0.7647   |
+| 0.6895        | 4.0   | 3504 | 0.6963          | 0.7680   |
+| 0.6712        | 5.0   | 4380 | 0.6856          | 0.7664   |
+| 0.653         | 6.0   | 5256 | 0.6824          | 0.7670   |
 
 
 ### Framework versions