whisper-medium-pt-cv19-fleurs2-lr-wu

This model is a fine-tuned version of openai/whisper-medium on the fsicoli/cv19-fleurs default dataset in Portuguese. It achieves the following results on the evaluation set:

• Loss: 0.2108
• Wer: 0.0949

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: 6.25e-06
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• training_steps: 25000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.0339	2.2883	5000	0.1694	0.1025
0.0281	4.5767	10000	0.1852	0.1005
0.0139	6.8650	15000	0.2092	0.1002
0.0044	9.1533	20000	0.2087	0.0960
0.0055	11.4416	25000	0.2108	0.0949

Framework versions

• Transformers 4.45.0.dev0
• Pytorch 2.4.1
• Datasets 2.21.0
• Tokenizers 0.19.1