update model card README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,106 @@
|
|
| 1 |
-
---
|
| 2 |
-
license:
|
| 3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
tags:
|
| 4 |
+
- generated_from_trainer
|
| 5 |
+
model-index:
|
| 6 |
+
- name: wav2vec2-xlsr-1B-NPSC-NN
|
| 7 |
+
results: []
|
| 8 |
+
---
|
| 9 |
+
|
| 10 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 11 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 12 |
+
|
| 13 |
+
# wav2vec2-xlsr-1B-NPSC-NN
|
| 14 |
+
|
| 15 |
+
This model is a fine-tuned version of [facebook/wav2vec2-xls-r-1b](https://huggingface.co/facebook/wav2vec2-xls-r-1b) on the None dataset.
|
| 16 |
+
It achieves the following results on the evaluation set:
|
| 17 |
+
- Loss: 0.4567
|
| 18 |
+
- Wer: 0.1533
|
| 19 |
+
|
| 20 |
+
## Model description
|
| 21 |
+
|
| 22 |
+
More information needed
|
| 23 |
+
|
| 24 |
+
## Intended uses & limitations
|
| 25 |
+
|
| 26 |
+
More information needed
|
| 27 |
+
|
| 28 |
+
## Training and evaluation data
|
| 29 |
+
|
| 30 |
+
More information needed
|
| 31 |
+
|
| 32 |
+
## Training procedure
|
| 33 |
+
|
| 34 |
+
### Training hyperparameters
|
| 35 |
+
|
| 36 |
+
The following hyperparameters were used during training:
|
| 37 |
+
- learning_rate: 6e-05
|
| 38 |
+
- train_batch_size: 8
|
| 39 |
+
- eval_batch_size: 8
|
| 40 |
+
- seed: 42
|
| 41 |
+
- gradient_accumulation_steps: 2
|
| 42 |
+
- total_train_batch_size: 16
|
| 43 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 44 |
+
- lr_scheduler_type: linear
|
| 45 |
+
- lr_scheduler_warmup_steps: 2000
|
| 46 |
+
- num_epochs: 50.0
|
| 47 |
+
- mixed_precision_training: Native AMP
|
| 48 |
+
|
| 49 |
+
### Training results
|
| 50 |
+
|
| 51 |
+
| Training Loss | Epoch | Step | Validation Loss | Wer |
|
| 52 |
+
|:-------------:|:-----:|:-----:|:---------------:|:------:|
|
| 53 |
+
| 1.6894 | 1.08 | 500 | 1.2423 | 0.8619 |
|
| 54 |
+
| 0.7543 | 2.15 | 1000 | 0.5956 | 0.3817 |
|
| 55 |
+
| 0.5481 | 3.23 | 1500 | 0.5043 | 0.3246 |
|
| 56 |
+
| 0.4661 | 4.3 | 2000 | 0.4813 | 0.2793 |
|
| 57 |
+
| 0.3901 | 5.38 | 2500 | 0.4371 | 0.2592 |
|
| 58 |
+
| 0.3512 | 6.45 | 3000 | 0.4216 | 0.2458 |
|
| 59 |
+
| 0.3016 | 7.53 | 3500 | 0.3814 | 0.2257 |
|
| 60 |
+
| 0.278 | 8.6 | 4000 | 0.4151 | 0.2145 |
|
| 61 |
+
| 0.2435 | 9.68 | 4500 | 0.4816 | 0.2130 |
|
| 62 |
+
| 0.2122 | 10.75 | 5000 | 0.4489 | 0.2137 |
|
| 63 |
+
| 0.1949 | 11.83 | 5500 | 0.3978 | 0.2063 |
|
| 64 |
+
| 0.1929 | 12.9 | 6000 | 0.3823 | 0.2026 |
|
| 65 |
+
| 0.1757 | 13.98 | 6500 | 0.3409 | 0.1965 |
|
| 66 |
+
| 0.1771 | 15.05 | 7000 | 0.3844 | 0.1936 |
|
| 67 |
+
| 0.1452 | 16.13 | 7500 | 0.3749 | 0.1900 |
|
| 68 |
+
| 0.1341 | 17.2 | 8000 | 0.4407 | 0.2026 |
|
| 69 |
+
| 0.13 | 18.28 | 8500 | 0.4253 | 0.1883 |
|
| 70 |
+
| 0.1183 | 19.35 | 9000 | 0.4311 | 0.1880 |
|
| 71 |
+
| 0.118 | 20.43 | 9500 | 0.4431 | 0.1882 |
|
| 72 |
+
| 0.1123 | 21.51 | 10000 | 0.4753 | 0.1820 |
|
| 73 |
+
| 0.1037 | 22.58 | 10500 | 0.4087 | 0.1834 |
|
| 74 |
+
| 0.1066 | 23.66 | 11000 | 0.4151 | 0.1845 |
|
| 75 |
+
| 0.0977 | 24.73 | 11500 | 0.4367 | 0.1783 |
|
| 76 |
+
| 0.0968 | 25.81 | 12000 | 0.4237 | 0.1756 |
|
| 77 |
+
| 0.0835 | 26.88 | 12500 | 0.4729 | 0.1781 |
|
| 78 |
+
| 0.0919 | 27.96 | 13000 | 0.4153 | 0.1701 |
|
| 79 |
+
| 0.0677 | 29.03 | 13500 | 0.4317 | 0.1693 |
|
| 80 |
+
| 0.0726 | 30.11 | 14000 | 0.4380 | 0.1736 |
|
| 81 |
+
| 0.066 | 31.18 | 14500 | 0.4384 | 0.1681 |
|
| 82 |
+
| 0.0713 | 32.26 | 15000 | 0.4215 | 0.1629 |
|
| 83 |
+
| 0.0605 | 33.33 | 15500 | 0.4574 | 0.1714 |
|
| 84 |
+
| 0.0632 | 34.41 | 16000 | 0.4343 | 0.1642 |
|
| 85 |
+
| 0.0567 | 35.48 | 16500 | 0.4231 | 0.1601 |
|
| 86 |
+
| 0.0556 | 36.56 | 17000 | 0.4404 | 0.1667 |
|
| 87 |
+
| 0.0426 | 37.63 | 17500 | 0.4459 | 0.1625 |
|
| 88 |
+
| 0.0445 | 38.71 | 18000 | 0.4484 | 0.1629 |
|
| 89 |
+
| 0.0463 | 39.78 | 18500 | 0.4508 | 0.1596 |
|
| 90 |
+
| 0.0448 | 40.86 | 19000 | 0.4395 | 0.1605 |
|
| 91 |
+
| 0.0434 | 41.94 | 19500 | 0.4490 | 0.1607 |
|
| 92 |
+
| 0.0347 | 43.01 | 20000 | 0.4772 | 0.1582 |
|
| 93 |
+
| 0.0332 | 44.09 | 20500 | 0.4729 | 0.1582 |
|
| 94 |
+
| 0.037 | 45.16 | 21000 | 0.4559 | 0.1573 |
|
| 95 |
+
| 0.0328 | 46.24 | 21500 | 0.4664 | 0.1560 |
|
| 96 |
+
| 0.0366 | 47.31 | 22000 | 0.4543 | 0.1543 |
|
| 97 |
+
| 0.0377 | 48.39 | 22500 | 0.4507 | 0.1560 |
|
| 98 |
+
| 0.0331 | 49.46 | 23000 | 0.4567 | 0.1533 |
|
| 99 |
+
|
| 100 |
+
|
| 101 |
+
### Framework versions
|
| 102 |
+
|
| 103 |
+
- Transformers 4.17.0.dev0
|
| 104 |
+
- Pytorch 1.10.1+cu102
|
| 105 |
+
- Datasets 1.18.2.dev0
|
| 106 |
+
- Tokenizers 0.11.0
|