update model card README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,100 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
tags:
|
| 4 |
+
- generated_from_trainer
|
| 5 |
+
model-index:
|
| 6 |
+
- name: wav2vec2-large-xlsr-53_english
|
| 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-large-xlsr-53_english
|
| 14 |
+
|
| 15 |
+
This model is a fine-tuned version of [facebook/wav2vec2-large-xlsr-53](https://huggingface.co/facebook/wav2vec2-large-xlsr-53) on the None dataset.
|
| 16 |
+
It achieves the following results on the evaluation set:
|
| 17 |
+
- Loss: 0.2620
|
| 18 |
+
- Wer: 0.1916
|
| 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: 0.0005
|
| 38 |
+
- train_batch_size: 32
|
| 39 |
+
- eval_batch_size: 8
|
| 40 |
+
- seed: 42
|
| 41 |
+
- gradient_accumulation_steps: 2
|
| 42 |
+
- total_train_batch_size: 64
|
| 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: 5
|
| 47 |
+
- mixed_precision_training: Native AMP
|
| 48 |
+
|
| 49 |
+
### Training results
|
| 50 |
+
|
| 51 |
+
| Training Loss | Epoch | Step | Validation Loss | Wer |
|
| 52 |
+
|:-------------:|:-----:|:-----:|:---------------:|:------:|
|
| 53 |
+
| 3.0506 | 0.12 | 250 | 3.0206 | 0.9999 |
|
| 54 |
+
| 1.4381 | 0.25 | 500 | 1.0267 | 0.6323 |
|
| 55 |
+
| 1.0903 | 0.37 | 750 | 0.5841 | 0.3704 |
|
| 56 |
+
| 1.0384 | 0.5 | 1000 | 0.5156 | 0.3348 |
|
| 57 |
+
| 0.9658 | 0.62 | 1250 | 0.4721 | 0.3221 |
|
| 58 |
+
| 0.9184 | 0.74 | 1500 | 0.4301 | 0.3213 |
|
| 59 |
+
| 0.8939 | 0.87 | 1750 | 0.4188 | 0.2884 |
|
| 60 |
+
| 0.9051 | 0.99 | 2000 | 0.3852 | 0.2807 |
|
| 61 |
+
| 0.563 | 1.12 | 2250 | 0.3752 | 0.2804 |
|
| 62 |
+
| 0.6122 | 1.24 | 2500 | 0.3745 | 0.2732 |
|
| 63 |
+
| 0.6213 | 1.36 | 2750 | 0.3671 | 0.2575 |
|
| 64 |
+
| 0.5839 | 1.49 | 3000 | 0.3560 | 0.2578 |
|
| 65 |
+
| 0.615 | 1.61 | 3250 | 0.3555 | 0.2536 |
|
| 66 |
+
| 0.5557 | 1.74 | 3500 | 0.3511 | 0.2485 |
|
| 67 |
+
| 0.5497 | 1.86 | 3750 | 0.3364 | 0.2425 |
|
| 68 |
+
| 0.5412 | 1.98 | 4000 | 0.3253 | 0.2418 |
|
| 69 |
+
| 0.2834 | 2.11 | 4250 | 0.3293 | 0.2322 |
|
| 70 |
+
| 0.2723 | 2.23 | 4500 | 0.3157 | 0.2322 |
|
| 71 |
+
| 0.2713 | 2.35 | 4750 | 0.3148 | 0.2304 |
|
| 72 |
+
| 0.2878 | 2.48 | 5000 | 0.3143 | 0.2286 |
|
| 73 |
+
| 0.2776 | 2.6 | 5250 | 0.3122 | 0.2250 |
|
| 74 |
+
| 0.2553 | 2.73 | 5500 | 0.3003 | 0.2234 |
|
| 75 |
+
| 0.278 | 2.85 | 5750 | 0.2973 | 0.2198 |
|
| 76 |
+
| 0.2445 | 2.97 | 6000 | 0.2938 | 0.2180 |
|
| 77 |
+
| 0.4361 | 3.1 | 6250 | 0.2914 | 0.2132 |
|
| 78 |
+
| 0.3979 | 3.22 | 6500 | 0.2916 | 0.2125 |
|
| 79 |
+
| 0.4221 | 3.35 | 6750 | 0.2879 | 0.2113 |
|
| 80 |
+
| 0.4051 | 3.47 | 7000 | 0.2819 | 0.2100 |
|
| 81 |
+
| 0.4218 | 3.59 | 7250 | 0.2812 | 0.2072 |
|
| 82 |
+
| 0.4201 | 3.72 | 7500 | 0.2772 | 0.2055 |
|
| 83 |
+
| 0.3515 | 3.84 | 7750 | 0.2747 | 0.2031 |
|
| 84 |
+
| 0.4021 | 3.97 | 8000 | 0.2702 | 0.2018 |
|
| 85 |
+
| 0.4304 | 4.09 | 8250 | 0.2721 | 0.2007 |
|
| 86 |
+
| 0.3923 | 4.21 | 8500 | 0.2689 | 0.1991 |
|
| 87 |
+
| 0.3824 | 4.34 | 8750 | 0.2692 | 0.1980 |
|
| 88 |
+
| 0.3743 | 4.46 | 9000 | 0.2718 | 0.1950 |
|
| 89 |
+
| 0.3771 | 4.59 | 9250 | 0.2653 | 0.1950 |
|
| 90 |
+
| 0.4048 | 4.71 | 9500 | 0.2649 | 0.1934 |
|
| 91 |
+
| 0.3539 | 4.83 | 9750 | 0.2638 | 0.1919 |
|
| 92 |
+
| 0.3498 | 4.96 | 10000 | 0.2620 | 0.1916 |
|
| 93 |
+
|
| 94 |
+
|
| 95 |
+
### Framework versions
|
| 96 |
+
|
| 97 |
+
- Transformers 4.11.3
|
| 98 |
+
- Pytorch 1.10.1+cu113
|
| 99 |
+
- Datasets 1.17.0
|
| 100 |
+
- Tokenizers 0.10.3
|