metadata
language:
- or
license: apache-2.0
tags:
- automatic-speech-recognition
- mozilla-foundation/common_voice_7_0
- generated_from_trainer
datasets:
- common_voice
model-index:
- name: wav2vec2-large-xls-r-300m-odia
results: []
wav2vec2-large-xls-r-300m-odia
This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on the MOZILLA-FOUNDATION/COMMON_VOICE_7_0 - OR dataset. It achieves the following results on the evaluation set:
python eval.py --model_id ./ --dataset mozilla-foundation/common_voice_7_0 --config as --split test --log_outputs
- WER: 0.7954545454545454
- CER: 0.32341269841269843
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: 7.5e-05
- train_batch_size: 16
- eval_batch_size: 16
- 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: 120.0
- mixed_precision_training: Native AMP
Training results
eval_loss | eval_wer | eval_runtime | eval_samples_per_second | eval_steps_per_second | epoch | |
---|---|---|---|---|---|---|
0 | 3.35224 | 0.998972 | 5.0475 | 22.189 | 1.387 | 29.41 |
1 | 1.33679 | 0.938335 | 5.0633 | 22.12 | 1.382 | 58.82 |
2 | 0.737202 | 0.957862 | 5.0913 | 21.998 | 1.375 | 88.24 |
3 | 0.658212 | 0.96814 | 5.0953 | 21.981 | 1.374 | 117.65 |
4 | 0.658 | 0.9712 | 5.0953 | 22.115 | 1.382 | 120 |
Framework versions
- Transformers 4.16.0.dev0
- Pytorch 1.10.1+cu102
- Datasets 1.17.1.dev0
- Tokenizers 0.11.0