End of training

README.md

@@ -22,7 +22,7 @@ model-index:
     metrics:
     - name: Wer
       type: wer
-      value: 0.14542967859585137
+      value: 0.23455664463186687
 ---
 
 <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -32,10 +32,10 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [openai/whisper-large-v2](https://huggingface.co/openai/whisper-large-v2) on the SwissDialDataset_ETH dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.2648
-- Wer Ortho: 0.2518
-- Wer: 0.1454
-- Cer: 0.0304
+- Loss: 0.2463
+- Wer Ortho: 0.3206
+- Wer: 0.2346
+- Cer: 0.0795
 
 ## Model description
 
@@ -55,9 +55,11 @@ More information needed
 
 The following hyperparameters were used during training:
 - learning_rate: 1e-05
-- train_batch_size: 16
-- eval_batch_size: 16
+- train_batch_size: 4
+- eval_batch_size: 4
 - seed: 42
+- gradient_accumulation_steps: 4
+- total_train_batch_size: 16
 - optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
 - lr_scheduler_type: constant_with_warmup
 - lr_scheduler_warmup_steps: 50
@@ -68,13 +70,13 @@ The following hyperparameters were used during training:
 
 | Training Loss | Epoch  | Step | Validation Loss | Wer Ortho | Wer    | Cer    |
 |:-------------:|:------:|:----:|:---------------:|:---------:|:------:|:------:|
-| 0.1387        | 1.2255 | 250  | 0.2670          | 0.2478    | 0.1523 | 0.0302 |
-| 0.0781        | 2.4510 | 500  | 0.2648          | 0.2518    | 0.1454 | 0.0304 |
+| 0.1296        | 1.2300 | 250  | 0.2512          | 0.3233    | 0.3987 | 0.2304 |
+| 0.0737        | 2.4600 | 500  | 0.2463          | 0.3206    | 0.2346 | 0.0795 |
 
 
 ### Framework versions
 
 - Transformers 4.46.3
 - Pytorch 2.5.1+cu121
-- Datasets 3.1.0
+- Datasets 3.2.0
 - Tokenizers 0.20.3