zuazo/whisper-base-gl-train

Whisper Base Galician

This model is a fine-tuned version of openai/whisper-base on the mozilla-foundation/common_voice_13_0 gl dataset. It achieves the following results on the evaluation set:

• Loss: 0.4494
• Wer: 17.9469

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: 2.5e-05
• train_batch_size: 128
• eval_batch_size: 64
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• training_steps: 5000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.3275	11.63	1000	0.4105	20.5409
0.1016	23.26	2000	0.4037	18.3862
0.0444	34.88	3000	0.4290	18.2859
0.0265	46.51	4000	0.4463	18.0144
0.0213	58.14	5000	0.4494	17.9469

Framework versions

• Transformers 4.37.2
• Pytorch 2.2.0+cu121
• Datasets 2.16.1
• Tokenizers 0.15.1