End of training

README.md

@@ -7,7 +7,7 @@ base_model: openai/whisper-small
 tags:
 - generated_from_trainer
 datasets:
-- mozilla-foundation/common_voice_17_0
+- mozilla-foundation/common_voice_11_0
 metrics:
 - wer
 model-index:
@@ -17,15 +17,15 @@ model-index:
       name: Automatic Speech Recognition
       type: automatic-speech-recognition
     dataset:
-      name: Common Voice 17.0
-      type: mozilla-foundation/common_voice_17_0
+      name: Common Voice 11.0
+      type: mozilla-foundation/common_voice_11_0
       config: ur
       split: test
       args: 'config: ur, split: test'
     metrics:
     - name: Wer
       type: wer
-      value: 92.16802838986547
+      value: 36.7003367003367
 ---
 
 <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -33,10 +33,10 @@ should probably proofread and complete it, then remove this comment. -->
 
 # Whisper Small Ur - Muhammad Abdullah
 
-This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the Common Voice 17.0 dataset.
+This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the Common Voice 11.0 dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.6097
-- Wer: 92.1680
+- Loss: 0.5832
+- Wer: 36.7003
 
 ## Model description
 
@@ -55,22 +55,24 @@ More information needed
 ### Training hyperparameters
 
 The following hyperparameters were used during training:
-- learning_rate: 5e-06
+- learning_rate: 1e-05
 - train_batch_size: 16
 - eval_batch_size: 8
 - seed: 42
+- gradient_accumulation_steps: 2
+- total_train_batch_size: 32
 - optimizer: Use 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: 100
-- training_steps: 500
+- lr_scheduler_warmup_steps: 50
+- training_steps: 400
 - mixed_precision_training: Native AMP
 
 ### Training results
 
-| Training Loss | Epoch  | Step | Validation Loss | Wer      |
-|:-------------:|:------:|:----:|:---------------:|:--------:|
-| 0.6086        | 0.7440 | 250  | 0.6541          | 102.7675 |
-| 0.4791        | 1.4881 | 500  | 0.6097          | 92.1680  |
+| Training Loss | Epoch  | Step | Validation Loss | Wer     |
+|:-------------:|:------:|:----:|:---------------:|:-------:|
+| 0.7572        | 1.5405 | 200  | 0.6012          | 38.1245 |
+| 0.4566        | 3.0772 | 400  | 0.5832          | 36.7003 |
 
 
 ### Framework versions

generation_config.json

@@ -160,6 +160,7 @@
     "<|yo|>": 50325,
     "<|zh|>": 50260
   },
+  "language": "ur",
   "max_initial_timestamp_index": 50,
   "max_length": 448,
   "no_timestamps_token_id": 50363,
@@ -254,6 +255,7 @@
     50361,
     50362
   ],
+  "task": "transcribe",
   "task_to_id": {
     "transcribe": 50359,
     "translate": 50358