harupurito/whisper-small-mar_cv

Whisper Small Mar - Harpreet Singh Anand

This model is a fine-tuned version of openai/whisper-small on the Common Voice 11.0 dataset. It achieves the following results on the evaluation set:

• Loss: 0.4596
• Wer: 43.3628

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
0.0658	4.07	1000	0.2928	46.3542
0.004	8.13	2000	0.3973	44.7295
0.0004	12.2	3000	0.4406	43.5046
0.0002	16.26	4000	0.4596	43.3628

Framework versions

• Transformers 4.39.2
• Pytorch 2.1.1+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2