Jerry02/whisper-tiny_to_indian_accent

Whisper tiny Indian

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

• Loss: 0.2640
• Wer: 13.0422

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-06
• train_batch_size: 4
• eval_batch_size: 1
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• training_steps: 4000

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.3405	2.3310	1000	0.2964	13.7801
0.3011	4.6620	2000	0.2720	13.1476
0.1913	6.9930	3000	0.2653	13.0572
0.1825	9.3240	4000	0.2640	13.0422

Framework versions

• Transformers 4.50.0.dev0
• Pytorch 2.5.1+cu124
• Datasets 3.3.2
• Tokenizers 0.21.0