rishabbahal/whisper-base-hi

Whisper Base Hi - Full fine tuned

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

• Loss: 0.5202
• Wer: 46.0226

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: Adam with betas=(0.9,0.999) and epsilon=1e-08
• 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
0.2093	2.3641	1000	0.4135	48.6273
0.0898	4.7281	2000	0.4273	45.7038
0.0279	7.0922	3000	0.4794	45.6334
0.0166	9.4563	4000	0.5202	46.0226

Framework versions

• Transformers 4.45.2
• Pytorch 2.4.1+cu121
• Datasets 3.0.1
• Tokenizers 0.20.1