Downloads · 30 days
7
18% of all-time downloads
wav2vec2-base-timit-demo-google-colab
k-al/wav2vec2-base-timit-demo-google-colab
wav2vec2-base-timit-demo-google-colab is a automatic speech recognition model from k-al. Use it when you need speech turned into text. It is set up for transformers. The card lists the license as apache-2.0.
should probably proofread and complete it, then remove this comment. --
All-time downloads
40
Public
Repo size
755 MB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.bin378 MB · 100%
At a glance
- Task
- Automatic Speech Recognition
- Library
- transformers
- License
- apache-2.0
- Model type
- wav2vec2
- Access
- Public
- Created
- Sep 3, 2022
- Updated
- Sep 3, 2022
- SHA
- 29c4448d
Model card
From the Hugging Face model README
<!-- 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-base-timit-demo-google-colab
This model is a fine-tuned version of facebook/wav2vec2-base on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.5167
- Wer: 0.3371
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: 0.0001
- train_batch_size: 8
- 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: 1000
- num_epochs: 30
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 3.5326 | 1.0 | 500 | 1.8362 | 1.0134 |
| 0.8886 | 2.01 | 1000 | 0.5096 | 0.5255 |
| 0.4265 | 3.01 | 1500 | 0.4255 | 0.4577 |
| 0.2955 | 4.02 | 2000 | 0.4041 | 0.4206 |
| 0.2251 | 5.02 | 2500 | 0.4508 | 0.4095 |
| 0.1967 | 6.02 | 3000 | 0.4277 | 0.3969 |
| 0.1651 | 7.03 | 3500 | 0.4357 | 0.3846 |
| 0.1387 | 8.03 | 4000 | 0.4546 | 0.4123 |
| 0.1235 | 9.04 | 4500 | 0.4817 | 0.3940 |
| 0.1124 | 10.04 | 5000 | 0.5007 | 0.3855 |
| 0.1046 | 11.04 | 5500 | 0.4883 | 0.3884 |
| 0.1179 | 12.05 | 6000 | 0.4652 | 0.3798 |
| 0.0858 | 13.05 | 6500 | 0.4652 | 0.3867 |
| 0.0767 | 14.06 | 7000 | 0.4919 | 0.3722 |
| 0.0658 | 15.06 | 7500 | 0.4844 | 0.3670 |
| 0.0605 | 16.06 | 8000 | 0.5113 | 0.3667 |
| 0.0579 | 17.07 | 8500 | 0.5369 | 0.3617 |
| 0.0549 | 18.07 | 9000 | 0.5116 | 0.3635 |
| 0.0452 | 19.08 | 9500 | 0.5099 | 0.3525 |
| 0.0425 | 20.08 | 10000 | 0.5324 | 0.3566 |
| 0.0405 | 21.08 | 10500 | 0.4881 | 0.3537 |
| 0.0338 | 22.09 | 11000 | 0.5161 | 0.3544 |
| 0.0388 | 23.09 | 11500 | 0.5216 | 0.3544 |
| 0.0358 | 24.1 | 12000 | 0.4999 | 0.3428 |
| 0.0284 | 25.1 | 12500 | 0.4905 | 0.3414 |
| 0.0236 | 26.1 | 13000 | 0.5029 | 0.3406 |
| 0.0245 | 27.11 | 13500 | 0.5087 | 0.3411 |
| 0.0218 | 28.11 | 14000 | 0.5188 | 0.3361 |
| 0.0217 | 29.12 | 14500 | 0.5167 | 0.3371 |
Framework versions
- Transformers 4.17.0
- Pytorch 1.12.1+cu113
- Datasets 1.18.3
- Tokenizers 0.12.1
- Task
- Automatic Speech Recognition
- Library
- transformers
- Type
- wav2vec2
- License
- apache-2.0
- Created
- Sep 3, 2022
- Updated
- Sep 3, 2022