Dacavi/whisper-small-es

Whisper Small Es - Spanish

This model is a fine-tuned version of openai/whisper-small on the Common Voice 11.0 dataset. It achieves the following results on the evaluation set:

• Loss: 0.1798
• Wer: 13.3333

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: 16
• 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: 50
• training_steps: 1000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.6172	0.1	100	0.6200	107.3958
0.2709	0.21	200	0.3492	67.0833
0.2839	0.31	300	0.2959	40.7292
0.2876	0.41	400	0.2766	29.5833
0.2296	0.52	500	0.2375	17.3958
0.2649	0.62	600	0.2102	15.3125
0.2644	0.72	700	0.1957	17.3958
0.2384	0.82	800	0.1886	13.7500
0.2325	0.93	900	0.1811	13.6458
0.1374	1.03	1000	0.1798	13.3333

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.0+cu118
• Datasets 2.15.0
• Tokenizers 0.15.0