BA_Model_V3

README.md

@@ -3,6 +3,8 @@ license: apache-2.0
 base_model: openai/whisper-large-v2
 tags:
 - generated_from_trainer
+metrics:
+- wer
 model-index:
 - name: BA_Model_V3
   results: []
@@ -14,6 +16,10 @@ should probably proofread and complete it, then remove this comment. -->
 # BA_Model_V3
 
 This model is a fine-tuned version of [openai/whisper-large-v2](https://huggingface.co/openai/whisper-large-v2) on an unknown dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.4521
+- Wer: 22.8402
+- Cer: 12.5097
 
 ## Model description
 
@@ -44,6 +50,20 @@ The following hyperparameters were used during training:
 - num_epochs: 8
 - mixed_precision_training: Native AMP
 
+### Training results
+
+| Training Loss | Epoch | Step | Validation Loss | Wer     | Cer     |
+|:-------------:|:-----:|:----:|:---------------:|:-------:|:-------:|
+| 0.3437        | 1.0   | 278  | 0.3292          | 23.6377 | 13.0135 |
+| 0.1749        | 2.0   | 556  | 0.3337          | 22.6309 | 12.1392 |
+| 0.0823        | 3.0   | 834  | 0.3635          | 23.1891 | 12.9637 |
+| 0.0397        | 4.0   | 1112 | 0.3959          | 23.1559 | 12.5382 |
+| 0.0202        | 5.0   | 1390 | 0.4146          | 24.8771 | 13.9211 |
+| 0.0097        | 6.0   | 1668 | 0.4196          | 22.1192 | 12.0233 |
+| 0.0038        | 7.0   | 1946 | 0.4424          | 22.2455 | 12.3308 |
+| 0.0018        | 8.0   | 2224 | 0.4521          | 22.8402 | 12.5097 |
+
+
 ### Framework versions
 
 - Transformers 4.38.1