zuazo/whisper-small-ca

Whisper Small Catalan

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

• Loss: 0.2169
• Wer: 10.0252

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: 64
• eval_batch_size: 32
• 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

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.1708	1.1	1000	0.2494	12.1846
0.0421	3.09	2000	0.2458	11.2689
0.0761	5.09	3000	0.2340	10.9231
0.0928	7.08	4000	0.2150	10.0394
0.0504	9.08	5000	0.2169	10.0252

Framework versions

• Transformers 4.33.0.dev0
• Pytorch 2.0.1+cu117
• Datasets 2.14.4
• Tokenizers 0.13.3