Commit ·
fef7a26
1
Parent(s): 10df70d
update model card README.md
Browse files
README.md
CHANGED
|
@@ -15,6 +15,9 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 15 |
# model-1h
|
| 16 |
|
| 17 |
This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on the common_voice dataset.
|
|
|
|
|
|
|
|
|
|
| 18 |
|
| 19 |
## Model description
|
| 20 |
|
|
@@ -41,12 +44,38 @@ The following hyperparameters were used during training:
|
|
| 41 |
- total_train_batch_size: 32
|
| 42 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 43 |
- lr_scheduler_type: linear
|
| 44 |
-
- lr_scheduler_warmup_steps:
|
| 45 |
- num_epochs: 30
|
| 46 |
- mixed_precision_training: Native AMP
|
| 47 |
|
| 48 |
### Training results
|
| 49 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 50 |
|
| 51 |
|
| 52 |
### Framework versions
|
|
|
|
| 15 |
# model-1h
|
| 16 |
|
| 17 |
This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on the common_voice dataset.
|
| 18 |
+
It achieves the following results on the evaluation set:
|
| 19 |
+
- Loss: 1.8317
|
| 20 |
+
- Wer: 1.0
|
| 21 |
|
| 22 |
## Model description
|
| 23 |
|
|
|
|
| 44 |
- total_train_batch_size: 32
|
| 45 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 46 |
- lr_scheduler_type: linear
|
| 47 |
+
- lr_scheduler_warmup_steps: 5
|
| 48 |
- num_epochs: 30
|
| 49 |
- mixed_precision_training: Native AMP
|
| 50 |
|
| 51 |
### Training results
|
| 52 |
|
| 53 |
+
| Training Loss | Epoch | Step | Validation Loss | Wer |
|
| 54 |
+
|:-------------:|:-----:|:----:|:---------------:|:------:|
|
| 55 |
+
| 11.4106 | 1.24 | 10 | 7.1597 | 1.0 |
|
| 56 |
+
| 4.777 | 2.47 | 20 | 3.9782 | 1.0 |
|
| 57 |
+
| 3.6585 | 3.71 | 30 | 3.3961 | 1.0 |
|
| 58 |
+
| 3.3143 | 4.94 | 40 | 3.1481 | 1.0 |
|
| 59 |
+
| 3.3318 | 6.24 | 50 | 3.0596 | 1.0 |
|
| 60 |
+
| 3.1368 | 7.47 | 60 | 2.9751 | 1.0 |
|
| 61 |
+
| 3.1058 | 8.71 | 70 | 2.9510 | 1.0 |
|
| 62 |
+
| 3.0605 | 9.94 | 80 | 2.9479 | 1.0 |
|
| 63 |
+
| 3.2043 | 11.24 | 90 | 2.9270 | 1.0 |
|
| 64 |
+
| 3.0424 | 12.47 | 100 | 2.9349 | 1.0 |
|
| 65 |
+
| 3.0374 | 13.71 | 110 | 2.9316 | 1.0 |
|
| 66 |
+
| 3.0256 | 14.94 | 120 | 2.9165 | 1.0 |
|
| 67 |
+
| 3.1724 | 16.24 | 130 | 2.9076 | 1.0 |
|
| 68 |
+
| 3.0119 | 17.47 | 140 | 2.9034 | 1.0 |
|
| 69 |
+
| 2.9937 | 18.71 | 150 | 2.8812 | 1.0 |
|
| 70 |
+
| 2.9775 | 19.94 | 160 | 2.8674 | 1.0 |
|
| 71 |
+
| 3.0826 | 21.24 | 170 | 2.8147 | 1.0 |
|
| 72 |
+
| 2.8717 | 22.47 | 180 | 2.7212 | 1.0 |
|
| 73 |
+
| 2.7714 | 23.71 | 190 | 2.6149 | 0.9952 |
|
| 74 |
+
| 2.634 | 24.94 | 200 | 2.4611 | 0.9984 |
|
| 75 |
+
| 2.5637 | 26.24 | 210 | 2.2734 | 1.0 |
|
| 76 |
+
| 2.237 | 27.47 | 220 | 2.0705 | 1.0 |
|
| 77 |
+
| 2.0381 | 28.71 | 230 | 1.9216 | 1.0 |
|
| 78 |
+
| 1.8788 | 29.94 | 240 | 1.8317 | 1.0 |
|
| 79 |
|
| 80 |
|
| 81 |
### Framework versions
|