jlondonobo/whisper-large-v2-pt-v3

Whisper Large Portuguese

This model is a fine-tuned version of openai/whisper-large-v2 on the mozilla-foundation/common_voice_11_0 pt dataset. It achieves the following results on the evaluation set:

• Loss: 0.1503
• Wer: 4.8385

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-06
• train_batch_size: 32
• eval_batch_size: 16
• seed: 42
• distributed_type: multi-GPU
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 300
• training_steps: 1500

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.1526	0.33	500	0.1588	4.9074
0.1046	1.3	1000	0.1510	4.8806
0.079	2.28	1500	0.1503	4.8385

Framework versions

• Transformers 4.26.0.dev0
• Pytorch 1.13.1+cu117
• Datasets 2.7.1.dev0
• Tokenizers 0.13.2