swahili

This model is a fine-tuned version of openai/whisper-small on the Common Voice 15.0 dataset. It achieves the following results on the evaluation set:

• Loss: 0.3824
• Wer: 36.4979

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: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• training_steps: 4000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.6697	0.1129	500	0.7159	64.0293
0.4719	0.2258	1000	0.5437	50.6878
0.4218	0.3388	1500	0.4773	45.0904
0.3896	0.4517	2000	0.4405	41.5501
0.3721	0.5646	2500	0.4173	39.9865
0.3386	0.6775	3000	0.3996	37.9094
0.3414	0.7904	3500	0.3883	37.3082
0.3078	0.9033	4000	0.3824	36.4979

Framework versions

• Transformers 4.40.0
• Pytorch 2.2.2+cu118
• Datasets 2.19.0
• Tokenizers 0.19.1