Edit model card

bert_base_tcm_0.7

This model is a fine-tuned version of neuralmind/bert-base-portuguese-cased on an unknown dataset. It achieves the following results on the evaluation set:

  • Loss: 0.0128
  • Criterio Julgamento Precision: 0.8235
  • Criterio Julgamento Recall: 0.9032
  • Criterio Julgamento F1: 0.8615
  • Criterio Julgamento Number: 93
  • Data Sessao Precision: 0.7324
  • Data Sessao Recall: 0.9286
  • Data Sessao F1: 0.8189
  • Data Sessao Number: 56
  • Modalidade Licitacao Precision: 0.9415
  • Modalidade Licitacao Recall: 0.9769
  • Modalidade Licitacao F1: 0.9589
  • Modalidade Licitacao Number: 346
  • Numero Exercicio Precision: 0.9486
  • Numero Exercicio Recall: 0.9486
  • Numero Exercicio F1: 0.9486
  • Numero Exercicio Number: 175
  • Objeto Licitacao Precision: 0.5352
  • Objeto Licitacao Recall: 0.6909
  • Objeto Licitacao F1: 0.6032
  • Objeto Licitacao Number: 55
  • Valor Objeto Precision: 0.8
  • Valor Objeto Recall: 0.8649
  • Valor Objeto F1: 0.8312
  • Valor Objeto Number: 37
  • Overall Precision: 0.8680
  • Overall Recall: 0.9318
  • Overall F1: 0.8987
  • Overall Accuracy: 0.9966

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: 8
  • eval_batch_size: 8
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 10.0

Training results

Training Loss Epoch Step Validation Loss Criterio Julgamento Precision Criterio Julgamento Recall Criterio Julgamento F1 Criterio Julgamento Number Data Sessao Precision Data Sessao Recall Data Sessao F1 Data Sessao Number Modalidade Licitacao Precision Modalidade Licitacao Recall Modalidade Licitacao F1 Modalidade Licitacao Number Numero Exercicio Precision Numero Exercicio Recall Numero Exercicio F1 Numero Exercicio Number Objeto Licitacao Precision Objeto Licitacao Recall Objeto Licitacao F1 Objeto Licitacao Number Valor Objeto Precision Valor Objeto Recall Valor Objeto F1 Valor Objeto Number Overall Precision Overall Recall Overall F1 Overall Accuracy
0.0267 1.0 2332 0.0175 0.8333 0.9140 0.8718 93 0.6825 0.7679 0.7227 56 0.9342 0.9855 0.9592 346 0.9194 0.9771 0.9474 175 0.4154 0.4909 0.45 55 0.5 0.7568 0.6022 37 0.8303 0.9121 0.8693 0.9954
0.0211 2.0 4664 0.0158 0.7154 0.9462 0.8148 93 0.7812 0.8929 0.8333 56 0.9319 0.9884 0.9593 346 0.9605 0.9714 0.9659 175 0.4 0.6545 0.4966 55 0.8293 0.9189 0.8718 37 0.8353 0.9449 0.8867 0.9956
0.0127 3.0 6996 0.0157 0.8218 0.8925 0.8557 93 0.8254 0.9286 0.8739 56 0.9522 0.9798 0.9658 346 0.96 0.96 0.96 175 0.5735 0.7091 0.6341 55 0.6857 0.6486 0.6667 37 0.8835 0.9252 0.9038 0.9957
0.0074 4.0 9328 0.0128 0.8235 0.9032 0.8615 93 0.7324 0.9286 0.8189 56 0.9415 0.9769 0.9589 346 0.9486 0.9486 0.9486 175 0.5352 0.6909 0.6032 55 0.8 0.8649 0.8312 37 0.8680 0.9318 0.8987 0.9966
0.0065 5.0 11660 0.0177 0.8113 0.9247 0.8643 93 0.675 0.9643 0.7941 56 0.9444 0.9827 0.9632 346 0.9392 0.9714 0.9551 175 0.5075 0.6182 0.5574 55 0.7674 0.8919 0.825 37 0.8566 0.9409 0.8968 0.9958
0.005 6.0 13992 0.0161 0.8485 0.9032 0.875 93 0.7164 0.8571 0.7805 56 0.9496 0.9798 0.9644 346 0.9556 0.9829 0.9690 175 0.6290 0.7091 0.6667 55 0.8108 0.8108 0.8108 37 0.8878 0.9344 0.9105 0.9967
0.0039 7.0 16324 0.0185 0.8925 0.8925 0.8925 93 0.7812 0.8929 0.8333 56 0.9602 0.9769 0.9685 346 0.9607 0.9771 0.9688 175 0.5224 0.6364 0.5738 55 0.8378 0.8378 0.8378 37 0.8951 0.9291 0.9118 0.9966
0.0035 8.0 18656 0.0188 0.8431 0.9247 0.8821 93 0.7903 0.875 0.8305 56 0.9571 0.9682 0.9626 346 0.9605 0.9714 0.9659 175 0.6981 0.6727 0.6852 55 0.8462 0.8919 0.8684 37 0.9068 0.9318 0.9191 0.9969
0.0017 9.0 20988 0.0207 0.8529 0.9355 0.8923 93 0.7727 0.9107 0.8361 56 0.9630 0.9769 0.9699 346 0.9605 0.9714 0.9659 175 0.7143 0.6364 0.6731 55 0.8462 0.8919 0.8684 37 0.9107 0.9370 0.9237 0.9968
0.002 10.0 23320 0.0191 0.8614 0.9355 0.8969 93 0.7647 0.9286 0.8387 56 0.9549 0.9798 0.9672 346 0.9553 0.9771 0.9661 175 0.6167 0.6727 0.6435 55 0.825 0.8919 0.8571 37 0.8954 0.9436 0.9188 0.9968

Framework versions

  • Transformers 4.21.0.dev0
  • Pytorch 1.11.0+cu113
  • Datasets 2.3.2
  • Tokenizers 0.12.1
Downloads last month
2
Hosted inference API
Token Classification
Examples
Examples
This model can be loaded on the Inference API on-demand.