Mark-X/whisper-small-or

Whisper Small Or- Mudit Sharma

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

• Loss: 0.3656
• Wer: 52.4265

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: 5000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.0192	7.8125	1000	0.2476	56.1209
0.0011	15.625	2000	0.2956	52.8128
0.0001	23.4375	3000	0.3320	52.6616
0.0	31.25	4000	0.3582	52.6280
0.0	39.0625	5000	0.3656	52.4265

Framework versions

• Transformers 4.44.0
• Pytorch 2.4.0
• Datasets 2.21.0
• Tokenizers 0.19.1