Singhamarjeet8130/whisper-medium-hi-ta

Whisper Medium Hi ta - Amarjeet

This model is a fine-tuned version of Singhamarjeet8130/whisper-medium-hi on the Common Voice 13.0 dataset. It achieves the following results on the evaluation set:

• Loss: 0.1794
• Wer: 37.3848

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: 16
• eval_batch_size: 8
• seed: 42
• optimizer: Use 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
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.133	0.2894	1000	0.2347	45.2937
0.1146	0.5787	2000	0.2040	41.4025
0.099	0.8681	3000	0.1835	38.8261
0.0652	1.1574	4000	0.1794	37.3848

Framework versions

• Transformers 4.47.0
• Pytorch 2.5.1+cu124
• Datasets 3.1.0
• Tokenizers 0.21.0