whisper-a-clp-ls-25

This model is a fine-tuned version of openai/whisper-small on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.0302
• Wer: 8.3857

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: 0.0004
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• 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: 132
• num_epochs: 11
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
No log	1.0	40	0.1805	27.4633
No log	2.0	80	0.2522	54.9266
1.2203	3.0	120	0.2866	59.5388
1.2203	4.0	160	0.1374	43.3962
0.1533	5.0	200	0.1516	46.1216
0.1533	6.0	240	0.1406	53.4591
0.1533	7.0	280	0.0704	17.8197
0.0561	8.0	320	0.0487	15.0943
0.0561	9.0	360	0.0395	11.7400
0.0225	10.0	400	0.0291	6.2893
0.0225	10.7342	429	0.0302	8.3857

Framework versions

• Transformers 4.47.0.dev0
• Pytorch 2.4.0
• Datasets 3.0.1
• Tokenizers 0.20.0