update model card README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,88 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
tags:
|
| 4 |
+
- generated_from_trainer
|
| 5 |
+
datasets:
|
| 6 |
+
- common_voice
|
| 7 |
+
model-index:
|
| 8 |
+
- name: wav2vec2-xlsr-1b-ru
|
| 9 |
+
results: []
|
| 10 |
+
---
|
| 11 |
+
|
| 12 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 13 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 14 |
+
|
| 15 |
+
# wav2vec2-xlsr-1b-ru
|
| 16 |
+
|
| 17 |
+
This model is a fine-tuned version of [facebook/wav2vec2-xls-r-1b](https://huggingface.co/facebook/wav2vec2-xls-r-1b) on the common_voice dataset.
|
| 18 |
+
It achieves the following results on the evaluation set:
|
| 19 |
+
- Loss: 0.1352
|
| 20 |
+
- Wer: 0.0971
|
| 21 |
+
|
| 22 |
+
## Model description
|
| 23 |
+
|
| 24 |
+
More information needed
|
| 25 |
+
|
| 26 |
+
## Intended uses & limitations
|
| 27 |
+
|
| 28 |
+
More information needed
|
| 29 |
+
|
| 30 |
+
## Training and evaluation data
|
| 31 |
+
|
| 32 |
+
More information needed
|
| 33 |
+
|
| 34 |
+
## Training procedure
|
| 35 |
+
|
| 36 |
+
### Training hyperparameters
|
| 37 |
+
|
| 38 |
+
The following hyperparameters were used during training:
|
| 39 |
+
- learning_rate: 5e-05
|
| 40 |
+
- train_batch_size: 32
|
| 41 |
+
- eval_batch_size: 8
|
| 42 |
+
- seed: 42
|
| 43 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 44 |
+
- lr_scheduler_type: linear
|
| 45 |
+
- lr_scheduler_warmup_steps: 500
|
| 46 |
+
- num_epochs: 10
|
| 47 |
+
- mixed_precision_training: Native AMP
|
| 48 |
+
|
| 49 |
+
### Training results
|
| 50 |
+
|
| 51 |
+
| Training Loss | Epoch | Step | Validation Loss | Wer |
|
| 52 |
+
|:-------------:|:-----:|:-----:|:---------------:|:------:|
|
| 53 |
+
| 0.5462 | 0.35 | 500 | 0.4027 | 0.3575 |
|
| 54 |
+
| 0.498 | 0.69 | 1000 | 0.2588 | 0.2513 |
|
| 55 |
+
| 0.4279 | 1.04 | 1500 | 0.2265 | 0.2204 |
|
| 56 |
+
| 0.4099 | 1.38 | 2000 | 0.2189 | 0.1979 |
|
| 57 |
+
| 0.4688 | 1.73 | 2500 | 0.2100 | 0.1920 |
|
| 58 |
+
| 0.2241 | 2.07 | 3000 | 0.1980 | 0.1767 |
|
| 59 |
+
| 0.2056 | 2.42 | 3500 | 0.2020 | 0.1683 |
|
| 60 |
+
| 0.3423 | 2.76 | 4000 | 0.1862 | 0.1606 |
|
| 61 |
+
| 0.2478 | 3.11 | 4500 | 0.1787 | 0.1563 |
|
| 62 |
+
| 0.3079 | 3.45 | 5000 | 0.1759 | 0.1555 |
|
| 63 |
+
| 0.2477 | 3.8 | 5500 | 0.1713 | 0.1423 |
|
| 64 |
+
| 0.1718 | 4.14 | 6000 | 0.1695 | 0.1391 |
|
| 65 |
+
| 0.1675 | 4.49 | 6500 | 0.1677 | 0.1372 |
|
| 66 |
+
| 0.1631 | 4.83 | 7000 | 0.1652 | 0.1333 |
|
| 67 |
+
| 0.1429 | 5.18 | 7500 | 0.1605 | 0.1308 |
|
| 68 |
+
| 0.1505 | 5.52 | 8000 | 0.1612 | 0.1245 |
|
| 69 |
+
| 0.1385 | 5.87 | 8500 | 0.1487 | 0.1225 |
|
| 70 |
+
| 0.1285 | 6.22 | 9000 | 0.1526 | 0.1201 |
|
| 71 |
+
| 0.1153 | 6.56 | 9500 | 0.1464 | 0.1172 |
|
| 72 |
+
| 0.1159 | 6.91 | 10000 | 0.1505 | 0.1143 |
|
| 73 |
+
| 0.1061 | 7.25 | 10500 | 0.1444 | 0.1106 |
|
| 74 |
+
| 0.1016 | 7.6 | 11000 | 0.1427 | 0.1075 |
|
| 75 |
+
| 0.1125 | 7.94 | 11500 | 0.1386 | 0.1045 |
|
| 76 |
+
| 0.0937 | 8.29 | 12000 | 0.1403 | 0.1022 |
|
| 77 |
+
| 0.1059 | 8.63 | 12500 | 0.1406 | 0.1022 |
|
| 78 |
+
| 0.0857 | 8.98 | 13000 | 0.1372 | 0.0992 |
|
| 79 |
+
| 0.0901 | 9.32 | 13500 | 0.1380 | 0.0977 |
|
| 80 |
+
| 0.0913 | 9.67 | 14000 | 0.1352 | 0.0971 |
|
| 81 |
+
|
| 82 |
+
|
| 83 |
+
### Framework versions
|
| 84 |
+
|
| 85 |
+
- Transformers 4.17.0.dev0
|
| 86 |
+
- Pytorch 1.10.2+cu102
|
| 87 |
+
- Datasets 1.18.3
|
| 88 |
+
- Tokenizers 0.11.0
|