wwwtwwwt/whisper-tiny-en-v1

Whisper Tiny En - Entertainment - Game Commentary

This model is a fine-tuned version of openai/whisper-tiny on the fineaudio-Entertainment-Game Commentary dataset. It achieves the following results on the evaluation set:

• Loss: 0.8809
• Wer: 44.2381

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: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• 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
0.8337	0.5984	1000	0.9700	53.8114
0.6236	1.1969	2000	0.9055	50.0491
0.6036	1.7953	3000	0.8853	46.1559
0.4801	2.3938	4000	0.8850	44.7257
0.4514	2.9922	5000	0.8809	44.2381

Framework versions

• Transformers 4.46.3
• Pytorch 2.4.0
• Datasets 3.1.0
• Tokenizers 0.20.0