whisper-small-fa

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: 0.9258
• Wer: 35.4973

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: 16
• 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: 100000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.0193	8.1103	20000	0.5349	36.7125
0.0046	16.2206	40000	0.6839	36.0033
0.0018	24.3309	60000	0.7936	36.2543
0.0003	32.4412	80000	0.8729	35.9406
0.0	40.5515	100000	0.9258	35.4973

Framework versions

• Transformers 4.45.1
• Pytorch 2.4.1+cu121
• Datasets 3.0.1
• Tokenizers 0.20.0