rajvs20/whisper-small-hi

Whisper Small Hi - Raj Vardhan

This model is a fine-tuned version of openai/whisper-small on the Common Voice 13 dataset. It achieves the following results on the evaluation set:

• Loss: 0.4428
• Wer Ortho: 33.1235
• Wer: 17.3730

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: 16
• 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: constant_with_warmup
• lr_scheduler_warmup_steps: 50
• training_steps: 4000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer Ortho	Wer
0.1225	1.1820	500	0.2934	36.5108	19.7346
0.0627	2.3641	1000	0.2926	34.1878	17.7126
0.0333	3.5461	1500	0.3187	32.8875	17.1898
0.0222	4.7281	2000	0.3486	32.9330	17.5093
0.0146	5.9102	2500	0.3722	33.2602	17.3954
0.0071	7.0922	3000	0.4006	32.5893	17.3596
0.0061	8.2742	3500	0.4318	33.2685	17.6032
0.0052	9.4563	4000	0.4428	33.1235	17.3730

Framework versions

• Transformers 4.48.1
• Pytorch 2.0.1+cu118
• Datasets 3.2.0
• Tokenizers 0.21.0