wav2vec2-xls-r-1b-medical-domain02

This model is a fine-tuned version of jonatasgrosman/wav2vec2-xls-r-1b-portuguese on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.7948
• Wer: 0.7625

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: 0.0003
• train_batch_size: 2
• eval_batch_size: 1
• seed: 42
• gradient_accumulation_steps: 8
• 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
• num_epochs: 40
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
No log	4.9689	100	3.0689	1.0
No log	9.9379	200	0.6458	0.6539
No log	14.9068	300	0.4059	0.4748
2.4261	19.8758	400	0.5375	0.5975
2.4261	24.8447	500	0.7920	0.7268
2.4261	29.8137	600	0.7915	0.7660
2.4261	34.7826	700	0.7948	0.7625
0.7229	39.7516	800	0.7948	0.7625

Framework versions

• Transformers 4.41.1
• Pytorch 2.1.2+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1