result
This model is a fine-tuned version of neuralmind/bert-large-portuguese-cased on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.7458
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: 5e-05
- train_batch_size: 4
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3.0
Training results
Framework versions
- Transformers 4.13.0.dev0
- Pytorch 1.10.0+cu102
- Datasets 1.16.1
- Tokenizers 0.10.3