WhisperTiny30hCommonVoice0.0Augment

This model is a fine-tuned version of OpenAI/whisper-tiny on the audiofolder dataset. It achieves the following results on the evaluation set:

• Loss: 0.8173
• Wer: 0.4007

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 OptimizerNames.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

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.7384	0.4405	500	0.9908	0.4775
0.6275	0.8811	1000	0.8960	0.4447
0.4919	1.3216	1500	0.8604	0.4225
0.4913	1.7621	2000	0.8450	0.4177
0.3667	2.2026	2500	0.8273	0.4090
0.3462	2.6432	3000	0.8362	0.4095
0.3239	3.0837	3500	0.8199	0.4061
0.3169	3.5242	4000	0.8202	0.4023
0.3123	3.9648	4500	0.8112	0.3996
0.2951	4.4053	5000	0.8173	0.4007

Framework versions

• Transformers 4.48.3
• Pytorch 2.6.0
• Datasets 3.3.0
• Tokenizers 0.21.0