lnxdx commited on
Commit
86d34a0
1 Parent(s): 3ca68aa

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +8 -6
README.md CHANGED
@@ -63,7 +63,7 @@ It achieves the following results:
63
  - WER on ShEMO dev set: 32.85
64
  - WER on Common Voice 13 test set: 19.21
65
 
66
- ## Evaluation results 🌡️
67
  | Checkpoint Name | WER on ShEMO dev set | WER on Common Voice 13 test set | Max :) |
68
  | :---------------------------------------------------------------------------------------------------------------: | :------: | :-------: | :---: |
69
  | [m3hrdadfi/wav2vec2-large-xlsr-persian-v3](https://huggingface.co/m3hrdadfi/wav2vec2-large-xlsr-persian-v3) | 46.55 | **17.43** | 46.55 |
@@ -73,7 +73,7 @@ It achieves the following results:
73
 
74
  As you can see, my model performs better in maximum case :D
75
 
76
- ## Training procedure
77
 
78
  #### Training hyperparameters
79
 
@@ -92,7 +92,7 @@ The following hyperparameters were used during training:
92
 
93
  You may need *gradient_accumulation* because you need more batch size.
94
 
95
- #### Training results
96
 
97
  | Training Loss | Epoch | Step | Validation Loss | Wer |
98
  |:-------------:|:-----:|:----:|:---------------:|:------:|
@@ -117,7 +117,7 @@ You may need *gradient_accumulation* because you need more batch size.
117
  | 0.8238 | 11.88 | 1900 | 0.6735 | 0.3297 |
118
  | 0.7618 | 12.5 | 2000 | 0.6728 | 0.3286 |
119
 
120
- #### Hyperparameter tuning
121
  Several models with differet hyperparameters were trained. The following figures show the training process for three of them.
122
  ![wer](wandb-wer.png)
123
  ![loss](wandb-loss.png)
@@ -182,5 +182,7 @@ Check out [this blog](https://huggingface.co/blog/fine-tune-xlsr-wav2vec2) for m
182
  ## Contact us 🤝
183
  If you have any technical question regarding the model, pretraining, code or publication, please create an issue in the repository. This is the *best* way to reach us.
184
 
185
- ## Citation ↩️
186
- *TO DO!*
 
 
 
63
  - WER on ShEMO dev set: 32.85
64
  - WER on Common Voice 13 test set: 19.21
65
 
66
+ ## Evaluation results 🧪
67
  | Checkpoint Name | WER on ShEMO dev set | WER on Common Voice 13 test set | Max :) |
68
  | :---------------------------------------------------------------------------------------------------------------: | :------: | :-------: | :---: |
69
  | [m3hrdadfi/wav2vec2-large-xlsr-persian-v3](https://huggingface.co/m3hrdadfi/wav2vec2-large-xlsr-persian-v3) | 46.55 | **17.43** | 46.55 |
 
73
 
74
  As you can see, my model performs better in maximum case :D
75
 
76
+ ## Training procedure 🏋️
77
 
78
  #### Training hyperparameters
79
 
 
92
 
93
  You may need *gradient_accumulation* because you need more batch size.
94
 
95
+ #### Training log
96
 
97
  | Training Loss | Epoch | Step | Validation Loss | Wer |
98
  |:-------------:|:-----:|:----:|:---------------:|:------:|
 
117
  | 0.8238 | 11.88 | 1900 | 0.6735 | 0.3297 |
118
  | 0.7618 | 12.5 | 2000 | 0.6728 | 0.3286 |
119
 
120
+ #### Hyperparameter tuning 🔧
121
  Several models with differet hyperparameters were trained. The following figures show the training process for three of them.
122
  ![wer](wandb-wer.png)
123
  ![loss](wandb-loss.png)
 
182
  ## Contact us 🤝
183
  If you have any technical question regarding the model, pretraining, code or publication, please create an issue in the repository. This is the *best* way to reach us.
184
 
185
+ ## Citation 🖇
186
+ *TO DO!*
187
+
188
+ **Fine-tuned with ❤️ without ☕︎**