đ sammy786/wav2vec2-xlsr-dhivehi
This model is a fine - tuned version of facebook/wav2vec2-xls-r-1b on the MOZILLA - FOUNDATION/COMMON_VOICE_8_0 - dv dataset. It's designed for automatic speech recognition, offering a practical solution for converting speech to text in the relevant language.
⨠Features
- Based on the pre - trained "facebook/wav2vec2-xls-r-1b" model, fine - tuned for better performance on the specific dataset.
- Achieves specific results on the evaluation set, including loss and WER values.
đĻ Installation
No installation steps are provided in the original document, so this section is skipped.
đģ Usage Examples
No code examples are provided in the original document, so this section is skipped.
đ Documentation
Model description
This model is a fine - tuned version of "facebook/wav2vec2-xls-r-1b". The fine - tuning process adapts the pre - trained model to the specific characteristics of the MOZILLA - FOUNDATION/COMMON_VOICE_8_0 - dv dataset.
Intended uses & limitations
More information needed.
Training and evaluation data
The training data consists of Common voice Finnish train.tsv, dev.tsv and other.tsv.
Training procedure
For creating the train dataset, all possible datasets were appended and a 90 - 10 split was used.
Training hyperparameters
The following hyperparameters were used during training:
| Property |
Details |
| learning_rate |
0.000045637994662983496 |
| train_batch_size |
8 |
| eval_batch_size |
16 |
| seed |
13 |
| gradient_accumulation_steps |
4 |
| total_train_batch_size |
32 |
| optimizer |
Adam with betas=(0.9,0.999) and epsilon = 1e - 08 |
| lr_scheduler_type |
cosine_with_restarts |
| lr_scheduler_warmup_steps |
500 |
| num_epochs |
30 |
| mixed_precision_training |
Native AMP |
Training results
| Step |
Training Loss |
Validation Loss |
Wer |
| 200 |
4.883800 |
3.190218 |
1.000000 |
| 400 |
1.600100 |
0.497887 |
0.726159 |
| 600 |
0.928500 |
0.358781 |
0.603892 |
| 800 |
0.867900 |
0.309132 |
0.570786 |
| 1000 |
0.743100 |
0.309116 |
0.552954 |
| 1200 |
0.725100 |
0.266839 |
0.538378 |
| 1400 |
0.786200 |
0.259797 |
0.535897 |
| 1600 |
0.655700 |
0.245691 |
0.517290 |
| 1800 |
0.650500 |
0.246957 |
0.516204 |
| 2000 |
0.685500 |
0.234808 |
0.516204 |
| 2200 |
0.487100 |
0.228409 |
0.507753 |
| 2400 |
0.401300 |
0.221087 |
0.495968 |
| 2600 |
0.359300 |
0.212476 |
0.489301 |
| 2800 |
0.347300 |
0.204848 |
0.487750 |
| 3000 |
0.327000 |
0.203163 |
0.478756 |
| 3200 |
0.337100 |
0.210235 |
0.487595 |
| 3400 |
0.308900 |
0.201471 |
0.491316 |
| 3600 |
0.292600 |
0.192437 |
0.476120 |
| 3800 |
0.289600 |
0.198398 |
0.468445 |
| 4000 |
0.290200 |
0.193484 |
0.467204 |
| 4200 |
0.272600 |
0.193999 |
0.470150 |
| 4400 |
0.266700 |
0.187384 |
0.460769 |
| 4600 |
0.253800 |
0.187279 |
0.476663 |
| 4800 |
0.266400 |
0.197395 |
0.466817 |
| 5000 |
0.258000 |
0.188920 |
0.456660 |
| 5200 |
0.237200 |
0.180770 |
0.457358 |
| 5400 |
0.237900 |
0.178149 |
0.448287 |
| 5600 |
0.232600 |
0.179827 |
0.461002 |
| 5800 |
0.228500 |
0.182142 |
0.445185 |
| 6000 |
0.221000 |
0.173619 |
0.440688 |
| 6200 |
0.219500 |
0.172291 |
0.442859 |
| 6400 |
0.219400 |
0.173339 |
0.430609 |
| 6600 |
0.201900 |
0.177552 |
0.426423 |
| 6800 |
0.199000 |
0.173157 |
0.429834 |
| 7000 |
0.200000 |
0.166503 |
0.423709 |
| 7200 |
0.194600 |
0.171812 |
0.429834 |
| 7400 |
0.192100 |
0.164989 |
0.420530 |
| 7600 |
0.185000 |
0.168355 |
0.418825 |
| 7800 |
0.175100 |
0.168128 |
0.419290 |
| 8000 |
0.173500 |
0.167959 |
0.424950 |
| 8200 |
0.172200 |
0.173643 |
0.414793 |
| 8400 |
0.164200 |
0.167020 |
0.406342 |
| 8600 |
0.170800 |
0.168050 |
0.405334 |
| 8800 |
0.157900 |
0.164290 |
0.396573 |
| 9000 |
0.159900 |
0.163188 |
0.397426 |
| 9200 |
0.151700 |
0.164370 |
0.390991 |
| 9400 |
0.146600 |
0.165053 |
0.392852 |
| 9600 |
0.142200 |
0.164939 |
0.391844 |
| 9800 |
0.148300 |
0.164422 |
0.385719 |
| 10000 |
0.136200 |
0.166569 |
0.385951 |
| 10200 |
0.140700 |
0.161377 |
0.379594 |
| 10400 |
0.133300 |
0.165194 |
0.378276 |
| 10600 |
0.131300 |
0.164328 |
0.369205 |
| 10800 |
0.135500 |
0.160254 |
0.373236 |
| 11000 |
0.121100 |
0.163522 |
0.372693 |
Framework versions
- Transformers 4.16.0.dev0
- Pytorch 1.10.0+cu102
- Datasets 1.17.1.dev0
- Tokenizers 0.10.3
Evaluation Commands
To evaluate on mozilla-foundation/common_voice_8_0 with split test, you can use the following command:
python eval.py --model_id sammy786/wav2vec2-xlsr-dhivehi --dataset mozilla-foundation/common_voice_8_0 --config dv --split test
đ§ Technical Details
This section is skipped as the original document does not contain sufficient technical details.
đ License
The model is licensed under the Apache - 2.0 license.
%20--%3e%3cdefs%3e%3cstyle%3e%20.st0%20{%20fill:%20%23061b40;%20}%20.st1%20{%20fill:%20%23306af1;%20}%20.st2%20{%20fill:%20%235ce5cf;%20}%20%3c/style%3e%3c/defs%3e%3cg%3e%3cpath%20class='st0'%20d='M55,10.5h9v3h-9v9h12V7.5h-12v3ZM64,19.5h-6v-3h6v3Z'/%3e%3cpolygon%20class='st0'%20points='69%2016.5%2078%2016.5%2078%2019.5%2069%2019.5%2069%2022.5%2081%2022.5%2081%2013.5%2072%2013.5%2072%2010.5%2081%2010.5%2081%207.5%2069%207.5%2069%2016.5'/%3e%3cpolygon%20class='st0'%20points='95%2010.5%2095%207.5%2083%207.5%2083%2022.5%2095%2022.5%2095%2019.5%2086%2019.5%2086%2016.5%2095%2016.5%2095%2013.5%2086%2013.5%2086%2010.5%2095%2010.5'/%3e%3cpath%20class='st0'%20d='M40,1.5v21h11.6l1.4-1.4v-7.6h0c0,0-1.4-1.5-1.4-1.5l1.4-1.4V2.9l-1.4-1.4h-11.6ZM50,19.5h-7v-6h7v6ZM50,10.5h-7v-6h7v6Z'/%3e%3c/g%3e%3cpath%20class='st1'%20d='M23.1,24L14.7,4.8l-4.9,11.2h3.8l-1.8,4H3.3L12.1,0H2C.9,0,0,.9,0,2v20c0,1.1.9,2,2,2h21.1Z'/%3e%3cpath%20class='st2'%20d='M34,0h-16.8l10.6,24h6.2c1.1,0,2-.9,2-2V2C36,.9,35.1,0,34,0ZM32.5,20h-4V4h4v16Z'/%3e%3c/svg%3e)
