whisper-small-mental-health-luganda-v1

This model is a fine-tuned version of sulaimank/whisper-small-CV-Fleurs-lg-300 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.1746
• Wer: 0.3639
• Cer: 0.1724

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: 5e-06
• train_batch_size: 32
• eval_batch_size: 16
• 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: 900
• num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
1.6554	1.0	297	1.7738	0.4284	0.2022
1.1974	2.0	594	1.3231	0.4490	0.2255
0.7331	3.0	891	0.9783	0.4034	0.1954
0.4240	4.0	1188	0.8874	0.3919	0.1891
0.2443	5.0	1485	0.9081	0.3809	0.1809
0.1337	6.0	1782	0.9790	0.3855	0.1838
0.0679	7.0	2079	1.0396	0.3896	0.1885
0.0337	8.0	2376	1.0917	0.3737	0.1794
0.0207	9.0	2673	1.1229	0.3706	0.1786
0.0154	10.0	2970	1.1359	0.3660	0.1715
0.0121	11.0	3267	1.1516	0.3634	0.1717
0.0099	12.0	3564	1.1586	0.3676	0.1737
0.0081	13.0	3861	1.1671	0.3640	0.1739
0.0067	14.0	4158	1.1746	0.3639	0.1724

Framework versions

• Transformers 5.3.0
• Pytorch 2.10.0+cu130
• Datasets 3.0.0
• Tokenizers 0.22.2