Commit
·
f5a8d84
1
Parent(s):
8b4d6f3
update model card README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,110 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
tags:
|
| 4 |
+
- generated_from_trainer
|
| 5 |
+
datasets:
|
| 6 |
+
- common_voice
|
| 7 |
+
model-index:
|
| 8 |
+
- name: wav2vec2-large-xls-r-300m-hi-cv8-m3
|
| 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-large-xls-r-300m-hi-cv8-m3
|
| 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: 0.6510
|
| 20 |
+
- Wer: 0.3179
|
| 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: 0.000239
|
| 40 |
+
- train_batch_size: 16
|
| 41 |
+
- eval_batch_size: 8
|
| 42 |
+
- seed: 42
|
| 43 |
+
- gradient_accumulation_steps: 2
|
| 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: 2000
|
| 48 |
+
- num_epochs: 50
|
| 49 |
+
- mixed_precision_training: Native AMP
|
| 50 |
+
|
| 51 |
+
### Training results
|
| 52 |
+
|
| 53 |
+
| Training Loss | Epoch | Step | Validation Loss | Wer |
|
| 54 |
+
|:-------------:|:-----:|:----:|:---------------:|:------:|
|
| 55 |
+
| 12.5576 | 1.04 | 200 | 6.6594 | 1.0 |
|
| 56 |
+
| 4.4069 | 2.07 | 400 | 3.6011 | 1.0 |
|
| 57 |
+
| 3.4273 | 3.11 | 600 | 3.3370 | 1.0 |
|
| 58 |
+
| 2.1108 | 4.15 | 800 | 1.0641 | 0.6562 |
|
| 59 |
+
| 0.8817 | 5.18 | 1000 | 0.7178 | 0.5172 |
|
| 60 |
+
| 0.6508 | 6.22 | 1200 | 0.6612 | 0.4839 |
|
| 61 |
+
| 0.5524 | 7.25 | 1400 | 0.6458 | 0.4889 |
|
| 62 |
+
| 0.4992 | 8.29 | 1600 | 0.5791 | 0.4382 |
|
| 63 |
+
| 0.4669 | 9.33 | 1800 | 0.6039 | 0.4352 |
|
| 64 |
+
| 0.4441 | 10.36 | 2000 | 0.6276 | 0.4297 |
|
| 65 |
+
| 0.4172 | 11.4 | 2200 | 0.6183 | 0.4474 |
|
| 66 |
+
| 0.3872 | 12.44 | 2400 | 0.5886 | 0.4231 |
|
| 67 |
+
| 0.3692 | 13.47 | 2600 | 0.6448 | 0.4399 |
|
| 68 |
+
| 0.3385 | 14.51 | 2800 | 0.6344 | 0.4075 |
|
| 69 |
+
| 0.3246 | 15.54 | 3000 | 0.5896 | 0.4087 |
|
| 70 |
+
| 0.3026 | 16.58 | 3200 | 0.6158 | 0.4016 |
|
| 71 |
+
| 0.284 | 17.62 | 3400 | 0.6038 | 0.3906 |
|
| 72 |
+
| 0.2682 | 18.65 | 3600 | 0.6165 | 0.3900 |
|
| 73 |
+
| 0.2577 | 19.69 | 3800 | 0.5754 | 0.3805 |
|
| 74 |
+
| 0.2509 | 20.73 | 4000 | 0.6028 | 0.3925 |
|
| 75 |
+
| 0.2426 | 21.76 | 4200 | 0.6335 | 0.4138 |
|
| 76 |
+
| 0.2346 | 22.8 | 4400 | 0.6128 | 0.3870 |
|
| 77 |
+
| 0.2205 | 23.83 | 4600 | 0.6223 | 0.3831 |
|
| 78 |
+
| 0.2104 | 24.87 | 4800 | 0.6122 | 0.3781 |
|
| 79 |
+
| 0.1992 | 25.91 | 5000 | 0.6467 | 0.3792 |
|
| 80 |
+
| 0.1916 | 26.94 | 5200 | 0.6277 | 0.3636 |
|
| 81 |
+
| 0.1835 | 27.98 | 5400 | 0.6317 | 0.3773 |
|
| 82 |
+
| 0.1776 | 29.02 | 5600 | 0.6124 | 0.3614 |
|
| 83 |
+
| 0.1751 | 30.05 | 5800 | 0.6475 | 0.3628 |
|
| 84 |
+
| 0.1662 | 31.09 | 6000 | 0.6266 | 0.3504 |
|
| 85 |
+
| 0.1584 | 32.12 | 6200 | 0.6347 | 0.3532 |
|
| 86 |
+
| 0.1494 | 33.16 | 6400 | 0.6636 | 0.3491 |
|
| 87 |
+
| 0.1457 | 34.2 | 6600 | 0.6334 | 0.3507 |
|
| 88 |
+
| 0.1427 | 35.23 | 6800 | 0.6397 | 0.3442 |
|
| 89 |
+
| 0.1397 | 36.27 | 7000 | 0.6468 | 0.3496 |
|
| 90 |
+
| 0.1283 | 37.31 | 7200 | 0.6291 | 0.3416 |
|
| 91 |
+
| 0.1255 | 38.34 | 7400 | 0.6652 | 0.3461 |
|
| 92 |
+
| 0.1195 | 39.38 | 7600 | 0.6587 | 0.3342 |
|
| 93 |
+
| 0.1169 | 40.41 | 7800 | 0.6478 | 0.3319 |
|
| 94 |
+
| 0.1126 | 41.45 | 8000 | 0.6280 | 0.3291 |
|
| 95 |
+
| 0.1112 | 42.49 | 8200 | 0.6434 | 0.3290 |
|
| 96 |
+
| 0.1069 | 43.52 | 8400 | 0.6542 | 0.3268 |
|
| 97 |
+
| 0.1027 | 44.56 | 8600 | 0.6536 | 0.3239 |
|
| 98 |
+
| 0.0993 | 45.6 | 8800 | 0.6622 | 0.3257 |
|
| 99 |
+
| 0.0973 | 46.63 | 9000 | 0.6572 | 0.3192 |
|
| 100 |
+
| 0.0911 | 47.67 | 9200 | 0.6522 | 0.3175 |
|
| 101 |
+
| 0.0897 | 48.7 | 9400 | 0.6521 | 0.3200 |
|
| 102 |
+
| 0.0905 | 49.74 | 9600 | 0.6510 | 0.3179 |
|
| 103 |
+
|
| 104 |
+
|
| 105 |
+
### Framework versions
|
| 106 |
+
|
| 107 |
+
- Transformers 4.16.2
|
| 108 |
+
- Pytorch 1.10.0+cu111
|
| 109 |
+
- Datasets 1.18.3
|
| 110 |
+
- Tokenizers 0.11.0
|