End of training

README.md

@@ -1,10 +1,10 @@
 ---
-base_model: facebook/wav2vec2-large-960h
 license: apache-2.0
-metrics:
-- wer
+base_model: facebook/wav2vec2-large-960h
 tags:
 - generated_from_trainer
+metrics:
+- wer
 model-index:
 - name: wav2vec2-base-cv-demo-google-colab
   results: []
@@ -17,8 +17,8 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [facebook/wav2vec2-large-960h](https://huggingface.co/facebook/wav2vec2-large-960h) on an unknown dataset.
 It achieves the following results on the evaluation set:
-- Loss: 1.4229
-- Wer: 0.7800
+- Loss: 0.3575
+- Wer: 0.2805
 
 ## Model description
 
@@ -43,15 +43,18 @@ The following hyperparameters were used during training:
 - seed: 42
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: linear
-- lr_scheduler_warmup_steps: 500
-- num_epochs: 2
+- lr_scheduler_warmup_steps: 1000
+- num_epochs: 3
 - mixed_precision_training: Native AMP
 
 ### Training results
 
 | Training Loss | Epoch  | Step | Validation Loss | Wer    |
 |:-------------:|:------:|:----:|:---------------:|:------:|
-| 5.9165        | 1.1905 | 500  | 1.4229          | 0.7800 |
+| 8.951         | 0.7126 | 300  | 3.0660          | 1.0    |
+| 3.0514        | 1.4252 | 600  | 2.9228          | 1.0    |
+| 2.7598        | 2.1378 | 900  | 0.7960          | 0.5544 |
+| 0.7975        | 2.8504 | 1200 | 0.3575          | 0.2805 |
 
 
 ### Framework versions