File size: 9,010 Bytes
0da561b |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 |
---
license: apache-2.0
base_model: facebook/wav2vec2-large-xlsr-53
tags:
- generated_from_trainer
datasets:
- fleurs
metrics:
- wer
model-index:
- name: wav2vec2_fleurs
results:
- task:
name: Automatic Speech Recognition
type: automatic-speech-recognition
dataset:
name: fleurs
type: fleurs
config: ar_eg
split: test
args: ar_eg
metrics:
- name: Wer
type: wer
value: 0.3367091772943236
---
<!-- 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. -->
# wav2vec2_fleurs
This model is a fine-tuned version of [facebook/wav2vec2-large-xlsr-53](https://huggingface.co/facebook/wav2vec2-large-xlsr-53) on the fleurs dataset.
It achieves the following results on the evaluation set:
- Loss: 0.4033
- Wer: 0.3367
## 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: 4e-05
- train_batch_size: 4
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 100
- num_epochs: 20
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|:-------------:|:-----:|:-----:|:---------------:|:------:|
| 18.6221 | 0.17 | 100 | 10.1383 | 1.0 |
| 6.5078 | 0.33 | 200 | 4.0182 | 1.0 |
| 3.632 | 0.5 | 300 | 3.2678 | 1.0 |
| 3.2359 | 0.67 | 400 | 3.1984 | 1.0 |
| 3.2014 | 0.83 | 500 | 3.1752 | 1.0 |
| 3.1857 | 1.0 | 600 | 3.1671 | 1.0 |
| 3.1816 | 1.17 | 700 | 3.1657 | 1.0 |
| 3.1912 | 1.33 | 800 | 3.1570 | 1.0 |
| 3.186 | 1.5 | 900 | 3.1548 | 1.0 |
| 3.1554 | 1.67 | 1000 | 3.1478 | 1.0 |
| 3.1521 | 1.83 | 1100 | 3.1442 | 1.0 |
| 3.1584 | 2.0 | 1200 | 3.1369 | 1.0 |
| 3.1554 | 2.17 | 1300 | 3.1340 | 1.0 |
| 3.172 | 2.33 | 1400 | 3.1304 | 1.0 |
| 3.1479 | 2.5 | 1500 | 3.1303 | 1.0 |
| 3.1359 | 2.67 | 1600 | 3.0864 | 1.0 |
| 3.0757 | 2.83 | 1700 | 2.9191 | 1.0 |
| 2.8491 | 3.0 | 1800 | 2.5490 | 1.0 |
| 2.4969 | 3.17 | 1900 | 1.9998 | 0.9785 |
| 2.048 | 3.33 | 2000 | 1.5004 | 0.9297 |
| 1.7632 | 3.5 | 2100 | 1.2369 | 0.8613 |
| 1.5885 | 3.67 | 2200 | 1.0752 | 0.7953 |
| 1.3712 | 3.83 | 2300 | 0.9573 | 0.7519 |
| 1.2916 | 4.0 | 2400 | 0.9038 | 0.7089 |
| 1.2559 | 4.17 | 2500 | 0.8269 | 0.6853 |
| 1.1625 | 4.33 | 2600 | 0.7781 | 0.6539 |
| 1.1264 | 4.5 | 2700 | 0.7555 | 0.6337 |
| 1.032 | 4.67 | 2800 | 0.7215 | 0.6032 |
| 1.0592 | 4.83 | 2900 | 0.6883 | 0.5734 |
| 0.9682 | 5.0 | 3000 | 0.6657 | 0.5504 |
| 0.9851 | 5.17 | 3100 | 0.6518 | 0.5448 |
| 0.9515 | 5.33 | 3200 | 0.6382 | 0.5403 |
| 0.9009 | 5.5 | 3300 | 0.6226 | 0.5296 |
| 0.9048 | 5.67 | 3400 | 0.6123 | 0.5161 |
| 0.8882 | 5.83 | 3500 | 0.6047 | 0.5098 |
| 0.8749 | 6.0 | 3600 | 0.5909 | 0.5006 |
| 0.7939 | 6.17 | 3700 | 0.5804 | 0.4931 |
| 0.8363 | 6.33 | 3800 | 0.5744 | 0.4877 |
| 0.8605 | 6.5 | 3900 | 0.5776 | 0.4884 |
| 0.8358 | 6.67 | 4000 | 0.5497 | 0.4745 |
| 0.7744 | 6.83 | 4100 | 0.5549 | 0.4664 |
| 0.7867 | 7.0 | 4200 | 0.5429 | 0.4629 |
| 0.7166 | 7.17 | 4300 | 0.5306 | 0.4465 |
| 0.7347 | 7.33 | 4400 | 0.5363 | 0.4521 |
| 0.7173 | 7.5 | 4500 | 0.5289 | 0.4429 |
| 0.7653 | 7.67 | 4600 | 0.5240 | 0.4389 |
| 0.7388 | 7.83 | 4700 | 0.5062 | 0.4304 |
| 0.7326 | 8.0 | 4800 | 0.5073 | 0.4290 |
| 0.6622 | 8.17 | 4900 | 0.5049 | 0.4236 |
| 0.7495 | 8.33 | 5000 | 0.5094 | 0.4254 |
| 0.6898 | 8.5 | 5100 | 0.4874 | 0.4216 |
| 0.6664 | 8.67 | 5200 | 0.4948 | 0.4225 |
| 0.6783 | 8.83 | 5300 | 0.4879 | 0.4131 |
| 0.7205 | 9.0 | 5400 | 0.4751 | 0.4136 |
| 0.6182 | 9.17 | 5500 | 0.4795 | 0.4085 |
| 0.6895 | 9.33 | 5600 | 0.4730 | 0.4099 |
| 0.6503 | 9.5 | 5700 | 0.4713 | 0.4029 |
| 0.624 | 9.67 | 5800 | 0.4699 | 0.4024 |
| 0.6268 | 9.83 | 5900 | 0.4726 | 0.4069 |
| 0.6525 | 10.0 | 6000 | 0.4593 | 0.3953 |
| 0.6112 | 10.17 | 6100 | 0.4558 | 0.3922 |
| 0.657 | 10.33 | 6200 | 0.4621 | 0.3940 |
| 0.6445 | 10.5 | 6300 | 0.4579 | 0.3906 |
| 0.5869 | 10.67 | 6400 | 0.4548 | 0.3903 |
| 0.5855 | 10.83 | 6500 | 0.4433 | 0.3840 |
| 0.5538 | 11.0 | 6600 | 0.4514 | 0.3897 |
| 0.5599 | 11.17 | 6700 | 0.4403 | 0.3786 |
| 0.5691 | 11.33 | 6800 | 0.4411 | 0.3800 |
| 0.5731 | 11.5 | 6900 | 0.4396 | 0.3768 |
| 0.5707 | 11.67 | 7000 | 0.4492 | 0.3770 |
| 0.5504 | 11.83 | 7100 | 0.4391 | 0.3690 |
| 0.6058 | 12.0 | 7200 | 0.4344 | 0.3717 |
| 0.5676 | 12.17 | 7300 | 0.4354 | 0.3758 |
| 0.5684 | 12.33 | 7400 | 0.4351 | 0.3656 |
| 0.5404 | 12.5 | 7500 | 0.4324 | 0.3636 |
| 0.5504 | 12.67 | 7600 | 0.4313 | 0.3658 |
| 0.5596 | 12.83 | 7700 | 0.4268 | 0.3632 |
| 0.5246 | 13.0 | 7800 | 0.4316 | 0.3633 |
| 0.5441 | 13.17 | 7900 | 0.4233 | 0.3648 |
| 0.5318 | 13.33 | 8000 | 0.4260 | 0.3597 |
| 0.5116 | 13.5 | 8100 | 0.4279 | 0.3591 |
| 0.5299 | 13.67 | 8200 | 0.4233 | 0.3606 |
| 0.5519 | 13.83 | 8300 | 0.4166 | 0.3567 |
| 0.5452 | 14.0 | 8400 | 0.4233 | 0.3573 |
| 0.5111 | 14.17 | 8500 | 0.4203 | 0.3580 |
| 0.5365 | 14.33 | 8600 | 0.4163 | 0.3577 |
| 0.5023 | 14.5 | 8700 | 0.4135 | 0.3552 |
| 0.5189 | 14.67 | 8800 | 0.4133 | 0.3485 |
| 0.5492 | 14.83 | 8900 | 0.4133 | 0.3478 |
| 0.5128 | 15.0 | 9000 | 0.4114 | 0.3478 |
| 0.486 | 15.17 | 9100 | 0.4222 | 0.3472 |
| 0.5015 | 15.33 | 9200 | 0.4129 | 0.3515 |
| 0.4871 | 15.5 | 9300 | 0.4132 | 0.3430 |
| 0.5267 | 15.67 | 9400 | 0.4109 | 0.3481 |
| 0.4814 | 15.83 | 9500 | 0.4109 | 0.3461 |
| 0.4801 | 16.0 | 9600 | 0.4140 | 0.3453 |
| 0.4894 | 16.17 | 9700 | 0.4074 | 0.3433 |
| 0.4756 | 16.33 | 9800 | 0.4070 | 0.3410 |
| 0.4446 | 16.5 | 9900 | 0.4088 | 0.3412 |
| 0.4838 | 16.67 | 10000 | 0.4070 | 0.3407 |
| 0.5087 | 16.83 | 10100 | 0.4048 | 0.3422 |
| 0.4994 | 17.0 | 10200 | 0.4043 | 0.3442 |
| 0.5421 | 17.17 | 10300 | 0.4088 | 0.3483 |
| 0.489 | 17.33 | 10400 | 0.4097 | 0.3450 |
| 0.4618 | 17.5 | 10500 | 0.4077 | 0.3430 |
| 0.4734 | 17.67 | 10600 | 0.4028 | 0.3433 |
| 0.4882 | 17.83 | 10700 | 0.4040 | 0.3393 |
| 0.4804 | 18.0 | 10800 | 0.4045 | 0.3385 |
| 0.483 | 18.17 | 10900 | 0.4055 | 0.3366 |
| 0.4916 | 18.33 | 11000 | 0.4077 | 0.3375 |
| 0.4933 | 18.5 | 11100 | 0.4056 | 0.3365 |
| 0.4881 | 18.67 | 11200 | 0.4023 | 0.3375 |
| 0.4869 | 18.83 | 11300 | 0.4031 | 0.3378 |
| 0.4649 | 19.0 | 11400 | 0.4026 | 0.3382 |
| 0.4793 | 19.17 | 11500 | 0.4035 | 0.3376 |
| 0.5252 | 19.33 | 11600 | 0.4019 | 0.3375 |
| 0.4681 | 19.5 | 11700 | 0.4026 | 0.3382 |
| 0.4311 | 19.67 | 11800 | 0.4026 | 0.3368 |
| 0.4799 | 19.83 | 11900 | 0.4034 | 0.3372 |
| 0.4323 | 20.0 | 12000 | 0.4033 | 0.3367 |
### Framework versions
- Transformers 4.36.0.dev0
- Pytorch 2.1.0+cu118
- Datasets 2.15.0
- Tokenizers 0.15.0
|