<!-- 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. -->
Whisper Small Odia v1.2
This model is a fine-tuned version of chandan3007/Whisper-small-oriya on the mozilla-foundation/common_voice_11_0 or dataset. It achieves the following results on the evaluation set:
- Loss: 0.4804
- Wer: 30.5219
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: 1e-05
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.98) and epsilon=1e-06
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 200
- training_steps: 1000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.0041 | 12.01 | 250 | 0.4540 | 31.5092 |
| 0.0011 | 24.02 | 500 | 0.4804 | 30.5219 |
| 0.0004 | 37.01 | 750 | 0.5306 | 31.4810 |
| 0.0 | 49.02 | 1000 | 0.5506 | 32.7786 |
Framework versions
- Transformers 4.26.1
- Pytorch 1.11.0+cu102
- Datasets 2.10.0
- Tokenizers 0.13.2
