Viraj008/whisper-small-mr_v4

Whisper Small MR v4 - Viraj Patil

This model is a fine-tuned version of Viraj008/whisper-small-mr_v3 on the Common Voice 17.0 dataset. It achieves the following results on the evaluation set:

• Loss: 0.2132
• Wer: 34.4975

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: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• 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
0.1241	0.5355	1000	0.2290	40.3892
0.0681	1.0710	2000	0.2150	36.1037
0.0576	1.6064	3000	0.2081	35.2573
0.0399	2.1419	4000	0.2132	34.4975

Framework versions

• Transformers 4.45.1
• Pytorch 2.4.0
• Datasets 3.0.1
• Tokenizers 0.20.0