File size: 4,323 Bytes
72b99e8 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 |
---
license: apache-2.0
tags:
- generated_from_trainer
metrics:
- wer
model-index:
- name: apac_5sents_XLS-R_2_1e-6
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# apac_5sents_XLS-R_2_1e-6
This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 267.7123
- Wer: 1.0
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 1e-06
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 128
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- training_steps: 5000
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|:-------------:|:------:|:----:|:---------------:|:---:|
| 277.1571 | 5.54 | 100 | 267.0546 | 1.0 |
| 276.4313 | 11.11 | 200 | 261.9242 | 1.0 |
| 265.9415 | 16.65 | 300 | 250.8011 | 1.0 |
| 238.6939 | 22.22 | 400 | 203.5241 | 1.0 |
| 174.4503 | 27.76 | 500 | 144.1427 | 1.0 |
| 133.3724 | 33.32 | 600 | 115.6477 | 1.0 |
| 111.8372 | 38.86 | 700 | 101.5777 | 1.0 |
| 100.9788 | 44.43 | 800 | 93.3811 | 1.0 |
| 93.1826 | 49.97 | 900 | 88.1119 | 1.0 |
| 88.9464 | 55.54 | 1000 | 84.3092 | 1.0 |
| 85.3244 | 61.11 | 1100 | 81.3484 | 1.0 |
| 81.8194 | 66.65 | 1200 | 78.9506 | 1.0 |
| 80.4364 | 72.22 | 1300 | 76.9131 | 1.0 |
| 77.8947 | 77.76 | 1400 | 75.1301 | 1.0 |
| 76.6954 | 83.32 | 1500 | 73.5633 | 1.0 |
| 74.5509 | 88.86 | 1600 | 72.1455 | 1.0 |
| 73.6445 | 94.43 | 1700 | 70.8411 | 1.0 |
| 72.1876 | 99.97 | 1800 | 69.6471 | 1.0 |
| 71.3292 | 105.54 | 1900 | 68.5407 | 1.0 |
| 70.0705 | 111.11 | 2000 | 67.4980 | 1.0 |
| 68.7978 | 116.65 | 2100 | 66.5158 | 1.0 |
| 68.0348 | 122.22 | 2200 | 65.5796 | 1.0 |
| 66.8713 | 127.76 | 2300 | 64.6976 | 1.0 |
| 66.16 | 133.32 | 2400 | 63.8526 | 1.0 |
| 65.2028 | 138.86 | 2500 | 63.0523 | 1.0 |
| 64.6085 | 144.43 | 2600 | 62.2911 | 1.0 |
| 63.5373 | 149.97 | 2700 | 61.5736 | 1.0 |
| 63.2075 | 155.54 | 2800 | 60.8822 | 1.0 |
| 62.4467 | 161.11 | 2900 | 60.2360 | 1.0 |
| 61.4208 | 166.65 | 3000 | 59.6212 | 1.0 |
| 61.1933 | 172.22 | 3100 | 59.0409 | 1.0 |
| 60.4399 | 177.76 | 3200 | 58.4939 | 1.0 |
| 60.0438 | 183.32 | 3300 | 57.9786 | 1.0 |
| 59.1202 | 188.86 | 3400 | 57.4928 | 1.0 |
| 59.1382 | 194.43 | 3500 | 57.0438 | 1.0 |
| 58.3211 | 199.97 | 3600 | 56.6216 | 1.0 |
| 58.1737 | 205.54 | 3700 | 56.2324 | 1.0 |
| 57.8397 | 211.11 | 3800 | 55.8701 | 1.0 |
| 57.1469 | 216.65 | 3900 | 55.5402 | 1.0 |
| 57.0738 | 222.22 | 4000 | 55.2410 | 1.0 |
| 56.6554 | 227.76 | 4100 | 54.9684 | 1.0 |
| 56.6029 | 233.32 | 4200 | 54.7281 | 1.0 |
| 55.9671 | 238.86 | 4300 | 54.5120 | 1.0 |
| 56.0056 | 244.43 | 4400 | 54.3272 | 1.0 |
| 55.7506 | 249.97 | 4500 | 54.1668 | 1.0 |
| 55.7896 | 255.54 | 4600 | 54.0354 | 1.0 |
| 55.6986 | 261.11 | 4700 | 53.9389 | 1.0 |
| 55.3473 | 266.65 | 4800 | 53.8633 | 1.0 |
| 55.4997 | 272.22 | 4900 | 53.8202 | 1.0 |
| 55.2924 | 277.76 | 5000 | 53.8053 | 1.0 |
### Framework versions
- Transformers 4.26.1
- Pytorch 2.0.1+cu118
- Datasets 2.11.0
- Tokenizers 0.13.3
|