Hanhpt23/whisper-tiny-silvarmed

openai/whisper-tiny

This model is a fine-tuned version of openai/whisper-tiny on the Hanhpt23/SilvarMed dataset. It achieves the following results on the evaluation set:

• Loss: 0.1653
• Wer: 2.6729

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
0.0595	1.0	2438	0.1336	4.7314
0.0301	2.0	4876	0.1503	5.6790
0.0202	3.0	7314	0.1391	5.3653
0.0181	4.0	9752	0.1544	6.9011
0.0155	5.0	12190	0.1623	4.4047
0.0067	6.0	14628	0.1711	4.1890
0.0075	7.0	17066	0.1636	3.7577
0.0073	8.0	19504	0.1676	3.2218
0.005	9.0	21942	0.1756	3.5616
0.0003	10.0	24380	0.1668	3.2675
0.0023	11.0	26818	0.1702	3.2741
0.0025	12.0	29256	0.1662	3.0257
0.0002	13.0	31694	0.1692	2.9996
0.0	14.0	34132	0.1755	4.4569
0.0008	15.0	36570	0.1713	3.0192
0.0001	16.0	39008	0.1620	2.7839
0.0	17.0	41446	0.1718	2.7317
0.0	18.0	43884	0.1659	2.7970
0.0	19.0	46322	0.1653	2.6925
0.0	20.0	48760	0.1653	2.6729

Framework versions

• Transformers 4.41.1
• Pytorch 2.3.0
• Datasets 3.2.0
• Tokenizers 0.19.1