deepdml/whisper-tiny-ta-mix-norm-optim

Whisper Tiny ta

This model is a fine-tuned version of openai/whisper-tiny on the Common Voice 17.0 dataset. It achieves the following results on the evaluation set:

• Loss: 0.4798
• Wer: 75.4222
• Cer: 21.7188

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: 64
• eval_batch_size: 64
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.04
• training_steps: 5000

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
0.4252	0.2	1000	0.5015	76.9190	22.6354
0.4034	0.4	2000	0.4818	75.6283	21.8341
0.4708	1.1734	3000	0.4798	75.1697	21.4852
0.4435	1.3734	4000	0.4797	75.3011	21.6158
0.457	2.1468	5000	0.4798	75.4222	21.7188

Framework versions

• Transformers 4.42.0.dev0
• Pytorch 2.3.0+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1

Citation

Please cite the model using the following BibTeX entry:

@misc{deepdml/whisper-tiny-ta-mix-norm-opti,
      title={Fine-tuned Whisper tiny ASR model for speech recognition in Tamil},
      author={Jimenez, David},
      howpublished={\url{https://huggingface.co/deepdml/whisper-tiny-ta-mix-norm-opti}},
      year={2026}
    }