Edit model card

bertimbau-base-lener-br-finetuned-lener-br

This model is a fine-tuned version of Luciano/bertimbau-base-finetuned-lener-br on the lener_br dataset. It achieves the following results on the evaluation set:

  • Loss: nan
  • Precision: 0.8943
  • Recall: 0.8970
  • F1: 0.8956
  • Accuracy: 0.9696

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: 15

Training results

Training Loss Epoch Step Validation Loss Precision Recall F1 Accuracy
0.0678 1.0 1957 nan 0.8148 0.8882 0.8499 0.9689
0.0371 2.0 3914 nan 0.8347 0.9022 0.8671 0.9671
0.0242 3.0 5871 nan 0.8491 0.8905 0.8693 0.9716
0.0197 4.0 7828 nan 0.9014 0.8772 0.8892 0.9780
0.0135 5.0 9785 nan 0.8651 0.9060 0.8851 0.9765
0.013 6.0 11742 nan 0.8882 0.9054 0.8967 0.9767
0.0084 7.0 13699 nan 0.8559 0.9097 0.8820 0.9751
0.0069 8.0 15656 nan 0.8916 0.8828 0.8872 0.9696
0.0047 9.0 17613 nan 0.8964 0.8931 0.8948 0.9716
0.0028 10.0 19570 nan 0.8864 0.9047 0.8955 0.9691
0.0023 11.0 21527 nan 0.8860 0.9011 0.8935 0.9693
0.0009 12.0 23484 nan 0.8952 0.8987 0.8970 0.9686
0.0014 13.0 25441 nan 0.8929 0.8985 0.8957 0.9699
0.0025 14.0 27398 nan 0.8914 0.8981 0.8947 0.9700
0.001 15.0 29355 nan 0.8943 0.8970 0.8956 0.9696

Framework versions

  • Transformers 4.21.2
  • Pytorch 1.12.1+cu113
  • Datasets 2.4.0
  • Tokenizers 0.12.1
Downloads last month
76
Safetensors
Model size
108M params
Tensor type
I64
·
F32
·
Inference Examples
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.

Dataset used to train Luciano/bertimbau-base-lener-br-finetuned-lener-br

Evaluation results