flima/grupo1-whisper-small-Spanish

Whisper Small ESPAÑOL - GRUPO 1

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

• Loss: 0.2025
• Wer: 8.9958

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: Use OptimizerNames.ADAMW_TORCH_FUSED 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: 50
• training_steps: 400
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.0127	5.0	25	0.1709	7.7406
0.0083	10.0	50	0.1787	7.7406
0.0046	15.0	75	0.1832	8.1590
0.0012	20.0	100	0.1883	8.5774
0.0007	25.0	125	0.1903	8.1590
0.0003	30.0	150	0.1927	8.3682
0.0002	35.0	175	0.1948	8.3682
0.0002	40.0	200	0.1965	8.3682
0.0002	45.0	225	0.1978	8.3682
0.0001	50.0	250	0.1992	8.5774
0.0001	55.0	275	0.2001	8.7866
0.0001	60.0	300	0.2009	8.7866
0.0001	65.0	325	0.2016	8.9958
0.0001	70.0	350	0.2021	8.9958
0.0001	75.0	375	0.2024	8.9958
0.0001	80.0	400	0.2025	8.9958

Framework versions

• Transformers 5.12.1
• Pytorch 2.11.0+cu128
• Datasets 5.0.0
• Tokenizers 0.22.2