End of training

README.md

@@ -19,8 +19,8 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the Common Voice 16.1 dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.2113
-- Cer: 10.2912
+- Loss: 0.2284
+- Cer: 10.0577
 
 ## Model description
 
@@ -46,21 +46,22 @@ The following hyperparameters were used during training:
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: linear
 - lr_scheduler_warmup_steps: 500
-- training_steps: 3000
+- training_steps: 4000
 - mixed_precision_training: Native AMP
 
 ### Training results
 
 | Training Loss | Epoch  | Step | Validation Loss | Cer     |
 |:-------------:|:------:|:----:|:---------------:|:-------:|
-| 0.1043        | 1.3245 | 1000 | 0.2119          | 10.7080 |
-| 0.0308        | 2.6490 | 2000 | 0.2085          | 10.6226 |
-| 0.0085        | 3.9735 | 3000 | 0.2113          | 10.2912 |
+| 0.1053        | 1.3245 | 1000 | 0.2142          | 10.8687 |
+| 0.0322        | 2.6490 | 2000 | 0.2128          | 10.3917 |
+| 0.0086        | 3.9735 | 3000 | 0.2213          | 10.1155 |
+| 0.0023        | 5.2980 | 4000 | 0.2284          | 10.0577 |
 
 
 ### Framework versions
 
-- Transformers 4.40.0
+- Transformers 4.40.1
 - Pytorch 2.1.2
 - Datasets 2.19.0
 - Tokenizers 0.19.1

generation_config.json

@@ -252,5 +252,5 @@
     "transcribe": 50359,
     "translate": 50358
   },
-  "transformers_version": "4.40.0"
+  "transformers_version": "4.40.1"
 }