This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on the MOZILLA-FOUNDATION/COMMON_VOICE_8_0 - MN dataset. It achieves the following results on the evaluation set:
- Loss: 0.5502
- Wer: 0.4042
Training and evaluation data
Evaluation is conducted in Notebook, you can see within the repo "notebook_evaluation_wav2vec2_mn.ipynb"
Test WER without LM wer = 58.2171 % cer = 16.0670 %
Test WER using wer = 31.3919 % cer = 10.2565 %
How to use eval.py
huggingface-cli login #login to huggingface for getting auth token to access the common voice v8
#running with LM
python eval.py --model_id ayameRushia/wav2vec2-large-xls-r-300m-mn --dataset mozilla-foundation/common_voice_8_0 --config mn --split test
# running without LM
python eval.py --model_id ayameRushia/wav2vec2-large-xls-r-300m-mn --dataset mozilla-foundation/common_voice_8_0 --config mn --split test --greedy
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.0003
- train_batch_size: 32
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 64
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- num_epochs: 40.0
- mixed_precision_training: Native AMP
Training results
Training Loss | Epoch | Step | Validation Loss | Wer |
---|---|---|---|---|
No log | 6.35 | 400 | 0.9380 | 0.7902 |
3.2674 | 12.7 | 800 | 0.5794 | 0.5309 |
0.7531 | 19.05 | 1200 | 0.5749 | 0.4815 |
0.5382 | 25.4 | 1600 | 0.5530 | 0.4447 |
0.4293 | 31.75 | 2000 | 0.5709 | 0.4237 |
0.4293 | 38.1 | 2400 | 0.5476 | 0.4059 |
Framework versions
- Transformers 4.17.0.dev0
- Pytorch 1.10.1+cu102
- Datasets 1.18.3
- Tokenizers 0.11.0
- Downloads last month
- 20
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social
visibility and check back later, or deploy to Inference Endpoints (dedicated)
instead.
Dataset used to train ayameRushia/wav2vec2-large-xls-r-300m-mn
Evaluation results
- Test WER using LM on Common Voice 8self-reported31.392
- Test CER using LM on Common Voice 8self-reported10.257
- Test WER on Robust Speech Event - Dev Dataself-reported65.260
- Test WER on Robust Speech Event - Test Dataself-reported63.090