Hanhpt23/whisper-small-vietmed-v1

openai/whisper-small

This model is a fine-tuned version of openai/whisper-small on the pphuc25/VietMed-split-8-2 dataset. It achieves the following results on the evaluation set:

• Loss: 0.9216
• Wer: 21.8561

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: 0.0001
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 100
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.6191	1.0	569	0.6122	27.3513
0.416	2.0	1138	0.5613	23.9795
0.2528	3.0	1707	0.6226	24.4701
0.1453	4.0	2276	0.6960	24.5140
0.0692	5.0	2845	0.7635	24.2687
0.0573	6.0	3414	0.7980	23.7013
0.032	7.0	3983	0.8142	23.7452
0.024	8.0	4552	0.8477	23.4999
0.024	9.0	5121	0.8555	23.6830
0.0126	10.0	5690	0.8619	23.4889
0.0086	11.0	6259	0.8640	23.3608
0.0061	12.0	6828	0.8844	22.7384
0.004	13.0	7397	0.8794	22.4492
0.0019	14.0	7966	0.8946	22.3833
0.0037	15.0	8535	0.9154	22.1636
0.0003	16.0	9104	0.9130	21.9843
0.0015	17.0	9673	0.9345	21.9074
0.0002	18.0	10242	0.9144	21.8708
0.0001	19.0	10811	0.9198	21.8268
0.0002	20.0	11380	0.9216	21.8561

Framework versions

• Transformers 4.41.1
• Pytorch 2.3.0
• Datasets 2.19.1
• Tokenizers 0.19.1