Abdullah17/whisper-small-urdu

Whisper Small UR - Muhammad Abdullah

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: 0.9758
• Wer: 41.6987

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 1e-05
• train_batch_size: 10
• 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: 3500
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.0074	9.62	1000	0.8238	42.0345
0.0003	19.23	2000	0.9381	42.6583
0.0002	28.85	3000	0.9758	41.6987

Framework versions

• Transformers 4.25.0.dev0
• Pytorch 1.12.1+cu113
• Datasets 2.7.0
• Tokenizers 0.13.2