metadata
license: mit
tags:
- generated_from_trainer
metrics:
- precision
- recall
- f1
- accuracy
model_index:
- name: bert-large-pt-archive
results:
- task:
name: Token Classification
type: token-classification
metric:
name: Accuracy
type: accuracy
value: 0.9766762474673703
bert-large-pt-archive
This model is a fine-tuned version of neuralmind/bert-large-portuguese-cased on an unkown dataset. It achieves the following results on the evaluation set:
- Loss: 0.0869
- Precision: 0.9280
- Recall: 0.9541
- F1: 0.9409
- Accuracy: 0.9767
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: 2e-05
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 4
Training results
| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
|---|---|---|---|---|---|---|---|
| 0.0665 | 1.0 | 765 | 0.1020 | 0.8928 | 0.9566 | 0.9236 | 0.9696 |
| 0.0392 | 2.0 | 1530 | 0.0781 | 0.9229 | 0.9586 | 0.9404 | 0.9757 |
| 0.0201 | 3.0 | 2295 | 0.0809 | 0.9278 | 0.9550 | 0.9412 | 0.9767 |
| 0.0152 | 4.0 | 3060 | 0.0869 | 0.9280 | 0.9541 | 0.9409 | 0.9767 |
Framework versions
- Transformers 4.10.0.dev0
- Pytorch 1.9.0+cu111
- Datasets 1.10.2
- Tokenizers 0.10.3