End of training

README.md

@@ -15,16 +15,16 @@ model-index:
 - name: ASR_Whisper_Peripheral_Neuropathy
   results:
   - task:
-      type: automatic-speech-recognition
       name: Automatic Speech Recognition
+      type: automatic-speech-recognition
     dataset:
       name: ASR_Preprocess_Peripheral_Neuropathy_Dataset
       type: yoona-J/ASR_Preprocess_Peripheral_Neuropathy_Dataset
       args: 'config: ko, split: valid'
     metrics:
-    - type: wer
-      value: 76.72287073065284
-      name: Wer
+    - name: Wer
+      type: wer
+      value: 15.849056603773585
 ---
 
 <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -34,9 +34,9 @@ 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 ASR_Preprocess_Peripheral_Neuropathy_Dataset dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.1934
-- Cer: 75.6174
-- Wer: 76.7229
+- Loss: 0.1775
+- Cer: 11.0108
+- Wer: 15.8491
 
 ## Model description
 
@@ -55,10 +55,12 @@ More information needed
 ### Training hyperparameters
 
 The following hyperparameters were used during training:
-- learning_rate: 1e-05
-- train_batch_size: 16
+- learning_rate: 3e-05
+- train_batch_size: 8
 - eval_batch_size: 8
 - seed: 42
+- gradient_accumulation_steps: 2
+- total_train_batch_size: 16
 - 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: 950
@@ -69,15 +71,15 @@ The following hyperparameters were used during training:
 
 | Training Loss | Epoch  | Step | Validation Loss | Cer     | Wer     |
 |:-------------:|:------:|:----:|:---------------:|:-------:|:-------:|
-| 0.3801        | 0.8562 | 1000 | 0.3118          | 84.9390 | 84.7328 |
-| 0.1683        | 1.7123 | 2000 | 0.2247          | 77.5879 | 75.5705 |
-| 0.082         | 2.5685 | 3000 | 0.1964          | 73.5959 | 68.2089 |
-| 0.0383        | 3.4247 | 4000 | 0.1936          | 69.9375 | 73.9813 |
-| 0.0156        | 4.2808 | 5000 | 0.1871          | 84.2450 | 83.5409 |
-| 0.0088        | 5.1370 | 6000 | 0.1906          | 71.3926 | 74.5129 |
-| 0.0088        | 5.9932 | 7000 | 0.1907          | 70.9905 | 76.3242 |
-| 0.0035        | 6.8493 | 8000 | 0.1922          | 76.4170 | 78.2131 |
-| 0.0011        | 7.7055 | 9000 | 0.1934          | 75.6174 | 76.7229 |
+| 0.3642        | 0.8565 | 1000 | 0.3042          | 88.0089 | 85.6774 |
+| 0.1563        | 1.7126 | 2000 | 0.2370          | 56.9486 | 47.3565 |
+| 0.0783        | 2.5687 | 3000 | 0.2043          | 16.7586 | 23.7618 |
+| 0.0305        | 3.4248 | 4000 | 0.1935          | 21.0814 | 25.2930 |
+| 0.0164        | 4.2809 | 5000 | 0.1843          | 19.5680 | 21.6927 |
+| 0.01          | 5.1370 | 6000 | 0.1823          | 13.3842 | 17.4606 |
+| 0.0086        | 5.9936 | 7000 | 0.1792          | 22.1973 | 24.9455 |
+| 0.003         | 6.8497 | 8000 | 0.1787          | 12.8683 | 17.4494 |
+| 0.0016        | 7.7058 | 9000 | 0.1775          | 11.0108 | 15.8491 |
 
 
 ### Framework versions

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:36dbe7921244fa26c314f3799fdaa3b21cb7a124801f7750a8ee711eeb84370f
+oid sha256:1d7c47e1b1ca96f16ae209479ba7ab3b4352936933db9e5f7e2e1924ca8b123e
 size 966995080