Kwimp/speed_augmentation

Whisper Small

This model is a fine-tuned version of openai/whisper-small on the Speechocean762_CMUkids_Myst dataset. It achieves the following results on the evaluation set:

• Loss: 0.4051
• Wer: 18.0804

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: 32
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 64
• 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: 2048
• training_steps: 4000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
5.0887	1.7544	500	3.0167	17.7263
3.1919	3.5088	1000	1.8832	16.8511
1.7489	5.2632	1500	1.0185	16.4881
0.8121	7.0175	2000	0.4592	18.8554
0.6613	8.7719	2500	0.4208	19.2183
0.6501	10.5263	3000	0.4108	18.6839
0.5977	12.2807	3500	0.4063	18.0091
0.5667	14.0351	4000	0.4051	18.0804

Framework versions

• Transformers 5.8.1
• Pytorch 2.5.1+cu121
• Datasets 4.8.5
• Tokenizers 0.22.2