Downloads · 30 days
6
1% of all-time downloads
wav2vec2-large-xls-r-300m-he
mtz2110/wav2vec2-large-xls-r-300m-he
wav2vec2-large-xls-r-300m-he is a automatic speech recognition model from mtz2110. 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
461
Public
Repo size
35.3 GB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.bin1.3 GB · 100%
At a glance
- Task
- Automatic Speech Recognition
- Library
- transformers
- License
- apache-2.0
- Model type
- wav2vec2
- Access
- Public
- Created
- Nov 27, 2022
- Updated
- Nov 28, 2022
- SHA
- 3d6bfa88
Datasets
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-large-xls-r-300m-he
This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on the fleurs dataset. It achieves the following results on the evaluation set:
- Loss: inf
- Wer: 0.5954
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: 0.0003
- train_batch_size: 2
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 6
- total_train_batch_size: 12
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- num_epochs: 20
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 8.8899 | 0.99 | 200 | inf | 1.0 |
| 3.0802 | 1.98 | 400 | inf | 1.0 |
| 1.4275 | 2.97 | 600 | inf | 0.8155 |
| 0.8737 | 3.96 | 800 | inf | 0.7276 |
| 0.6503 | 4.95 | 1000 | inf | 0.6858 |
| 0.5176 | 5.94 | 1200 | inf | 0.6660 |
| 0.4084 | 6.93 | 1400 | inf | 0.6682 |
| 0.3469 | 7.92 | 1600 | inf | 0.6473 |
| 3.2485 | 6.67 | 1800 | inf | 1.0 |
| 0.6476 | 7.41 | 2000 | inf | 0.6574 |
| 0.3229 | 8.15 | 2200 | inf | 0.6499 |
| 0.2899 | 8.89 | 2400 | inf | 0.6376 |
| 0.26 | 9.63 | 2600 | inf | 0.6405 |
| 0.2038 | 10.37 | 2800 | inf | 0.6409 |
| 0.2158 | 11.11 | 3000 | inf | 0.6313 |
| 0.1892 | 11.85 | 3200 | inf | 0.6185 |
| 0.1611 | 12.59 | 3400 | inf | 0.6271 |
| 0.1584 | 13.33 | 3600 | inf | 0.6101 |
| 0.1443 | 14.07 | 3800 | inf | 0.6121 |
| 0.1353 | 14.81 | 4000 | inf | 0.6194 |
| 0.1109 | 15.56 | 4200 | inf | 0.6321 |
| 0.1116 | 16.3 | 4400 | inf | 0.6025 |
| 0.1054 | 17.04 | 4600 | inf | 0.6029 |
| 0.0966 | 17.78 | 4800 | inf | 0.6069 |
| 0.0824 | 18.52 | 5000 | inf | 0.5998 |
| 0.0812 | 19.26 | 5200 | inf | 0.5972 |
| 0.0749 | 20.0 | 5400 | inf | 0.5954 |
Framework versions
- Transformers 4.24.0
- Pytorch 1.12.1
- Datasets 2.6.1
- Tokenizers 0.13.1
- Task
- Automatic Speech Recognition
- Library
- transformers
- Type
- wav2vec2
- License
- apache-2.0
- Created
- Nov 27, 2022
- Updated
- Nov 28, 2022