metadata
base_model: facebook/wav2vec2-xls-r-300m
datasets:
- common_voice_13_0
license: apache-2.0
metrics:
- wer
tags:
- generated_from_trainer
model-index:
- name: wav2vec2-large-xls-r-300m-tamil-commonvoice
results:
- task:
type: automatic-speech-recognition
name: Automatic Speech Recognition
dataset:
name: common_voice_13_0
type: common_voice_13_0
config: ta
split: test
args: ta
metrics:
- type: wer
value: 1
name: Wer
wav2vec2-large-xls-r-300m-tamil-commonvoice
This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on the common_voice_13_0 dataset. It achieves the following results on the evaluation set:
- Loss: 7.9682
- 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: 0.0003
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- num_epochs: 30
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 6.0629 | 0.7737 | 400 | 1.5761 | 0.9986 |
| 0.6711 | 1.5474 | 800 | 0.5474 | 0.7253 |
| 0.437 | 2.3211 | 1200 | 0.4898 | 0.6689 |
| 0.3691 | 3.0948 | 1600 | 0.4760 | 0.6562 |
| 0.3942 | 3.8685 | 2000 | 0.8449 | 0.7908 |
| 1.3114 | 4.6422 | 2400 | 1.8169 | 0.9883 |
| 3.0292 | 5.4159 | 2800 | 3.3102 | 1.0 |
| 3.3769 | 6.1896 | 3200 | 3.4855 | 1.0 |
| 4.0469 | 6.9632 | 3600 | 5.2510 | 1.0 |
| 6.6565 | 7.7369 | 4000 | 7.9749 | 1.0 |
| 7.9329 | 8.5106 | 4400 | 7.9682 | 1.0 |
| 7.925 | 9.2843 | 4800 | 7.9682 | 1.0 |
| 7.9128 | 10.0580 | 5200 | 7.9682 | 1.0 |
| 7.9132 | 10.8317 | 5600 | 7.9682 | 1.0 |
| 7.9118 | 11.6054 | 6000 | 7.9682 | 1.0 |
| 7.8873 | 12.3791 | 6400 | 7.9682 | 1.0 |
| 7.9357 | 13.1528 | 6800 | 7.9682 | 1.0 |
| 7.9311 | 13.9265 | 7200 | 7.9682 | 1.0 |
| 7.9049 | 14.7002 | 7600 | 7.9682 | 1.0 |
| 7.9234 | 15.4739 | 8000 | 7.9682 | 1.0 |
| 7.9521 | 16.2476 | 8400 | 7.9682 | 1.0 |
| 7.8886 | 17.0213 | 8800 | 7.9682 | 1.0 |
| 7.8915 | 17.7950 | 9200 | 7.9682 | 1.0 |
| 7.9265 | 18.5687 | 9600 | 7.9682 | 1.0 |
| 7.9366 | 19.3424 | 10000 | 7.9682 | 1.0 |
| 7.8725 | 20.1161 | 10400 | 7.9682 | 1.0 |
| 7.9321 | 20.8897 | 10800 | 7.9682 | 1.0 |
| 7.9282 | 21.6634 | 11200 | 7.9682 | 1.0 |
| 7.9025 | 22.4371 | 11600 | 7.9682 | 1.0 |
| 7.8889 | 23.2108 | 12000 | 7.9682 | 1.0 |
| 7.9366 | 23.9845 | 12400 | 7.9682 | 1.0 |
| 7.9205 | 24.7582 | 12800 | 7.9682 | 1.0 |
| 7.8946 | 25.5319 | 13200 | 7.9682 | 1.0 |
| 7.9446 | 26.3056 | 13600 | 7.9682 | 1.0 |
| 7.8891 | 27.0793 | 14000 | 7.9682 | 1.0 |
| 7.9088 | 27.8530 | 14400 | 7.9682 | 1.0 |
| 7.9546 | 28.6267 | 14800 | 7.9682 | 1.0 |
| 7.8624 | 29.4004 | 15200 | 7.9682 | 1.0 |
Framework versions
- Transformers 4.40.1
- Pytorch 1.13.1+cu117
- Datasets 2.18.0
- Tokenizers 0.19.1