Edit model card

predict-perception-bert-cause-concept

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

  • Loss: 0.4044
  • Rmse: 0.6076
  • Rmse Cause::a Causata da un concetto astratto (es. gelosia): 0.6076
  • Mae: 0.4548
  • Mae Cause::a Causata da un concetto astratto (es. gelosia): 0.4548
  • R2: 0.5463
  • R2 Cause::a Causata da un concetto astratto (es. gelosia): 0.5463
  • Cos: 0.2174
  • Pair: 0.0
  • Rank: 0.5
  • Neighbors: 0.3931
  • Rsa: nan

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: 1e-05
  • train_batch_size: 20
  • eval_batch_size: 8
  • seed: 1996
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 30

Training results

Training Loss Epoch Step Validation Loss Rmse Rmse Cause::a Causata da un concetto astratto (es. gelosia) Mae Mae Cause::a Causata da un concetto astratto (es. gelosia) R2 R2 Cause::a Causata da un concetto astratto (es. gelosia) Cos Pair Rank Neighbors Rsa
1.08 1.0 15 0.9520 0.9323 0.9323 0.6560 0.6560 -0.0680 -0.0680 0.0435 0.0 0.5 0.3188 nan
0.9974 2.0 30 0.8621 0.8872 0.8872 0.5962 0.5962 0.0328 0.0328 0.1304 0.0 0.5 0.4066 nan
0.9337 3.0 45 0.9223 0.9176 0.9176 0.6608 0.6608 -0.0347 -0.0347 0.2174 0.0 0.5 0.3632 nan
0.966 4.0 60 0.8273 0.8691 0.8691 0.5874 0.5874 0.0719 0.0719 0.2174 0.0 0.5 0.3754 nan
0.8683 5.0 75 0.8741 0.8933 0.8933 0.6136 0.6136 0.0193 0.0193 0.2174 0.0 0.5 0.3529 nan
0.8522 6.0 90 0.7781 0.8428 0.8428 0.5732 0.5732 0.1271 0.1271 0.2174 0.0 0.5 0.4152 nan
0.7968 7.0 105 0.7257 0.8139 0.8139 0.5519 0.5519 0.1859 0.1859 0.2174 0.0 0.5 0.4152 nan
0.7166 8.0 120 0.7122 0.8064 0.8064 0.5792 0.5792 0.2010 0.2010 0.1304 0.0 0.5 0.3955 nan
0.6246 9.0 135 0.6771 0.7862 0.7862 0.5701 0.5701 0.2403 0.2403 0.0435 0.0 0.5 0.3955 nan
0.5205 10.0 150 0.6704 0.7823 0.7823 0.5735 0.5735 0.2479 0.2479 0.3913 0.0 0.5 0.4847 nan
0.4182 11.0 165 0.6852 0.7909 0.7909 0.5987 0.5987 0.2313 0.2313 0.3913 0.0 0.5 0.4847 nan
0.3984 12.0 180 0.6106 0.7466 0.7466 0.5696 0.5696 0.3150 0.3150 0.0435 0.0 0.5 0.2935 nan
0.3138 13.0 195 0.5867 0.7318 0.7318 0.5209 0.5209 0.3418 0.3418 0.2174 0.0 0.5 0.3119 nan
0.2323 14.0 210 0.5120 0.6837 0.6837 0.5007 0.5007 0.4256 0.4256 0.3043 0.0 0.5 0.3849 nan
0.2149 15.0 225 0.4789 0.6612 0.6612 0.4883 0.4883 0.4627 0.4627 0.3043 0.0 0.5 0.3849 nan
0.1753 16.0 240 0.4526 0.6428 0.6428 0.4775 0.4775 0.4922 0.4922 0.3043 0.0 0.5 0.3849 nan
0.1478 17.0 255 0.4383 0.6325 0.6325 0.4616 0.4616 0.5083 0.5083 0.2174 0.0 0.5 0.3931 nan
0.1289 18.0 270 0.4141 0.6148 0.6148 0.4478 0.4478 0.5355 0.5355 0.3043 0.0 0.5 0.3849 nan
0.1035 19.0 285 0.3952 0.6007 0.6007 0.4407 0.4407 0.5566 0.5566 0.3043 0.0 0.5 0.3849 nan
0.1087 20.0 300 0.4217 0.6205 0.6205 0.4505 0.4505 0.5269 0.5269 0.2174 0.0 0.5 0.3931 nan
0.1005 21.0 315 0.4065 0.6091 0.6091 0.4508 0.4508 0.5440 0.5440 0.2174 0.0 0.5 0.3931 nan
0.0868 22.0 330 0.3937 0.5995 0.5995 0.4470 0.4470 0.5584 0.5584 0.3043 0.0 0.5 0.3849 nan
0.0808 23.0 345 0.4132 0.6142 0.6142 0.4617 0.4617 0.5364 0.5364 0.2174 0.0 0.5 0.3931 nan
0.0737 24.0 360 0.4214 0.6203 0.6203 0.4659 0.4659 0.5272 0.5272 0.3043 0.0 0.5 0.4066 nan
0.0711 25.0 375 0.3863 0.5939 0.5939 0.4470 0.4470 0.5666 0.5666 0.3043 0.0 0.5 0.3849 nan
0.066 26.0 390 0.4353 0.6304 0.6304 0.4760 0.4760 0.5117 0.5117 0.2174 0.0 0.5 0.3931 nan
0.0681 27.0 405 0.4078 0.6101 0.6101 0.4612 0.4612 0.5426 0.5426 0.2174 0.0 0.5 0.3931 nan
0.0543 28.0 420 0.4118 0.6132 0.6132 0.4616 0.4616 0.5380 0.5380 0.2174 0.0 0.5 0.3931 nan
0.069 29.0 435 0.4041 0.6074 0.6074 0.4551 0.4551 0.5466 0.5466 0.2174 0.0 0.5 0.3931 nan
0.0604 30.0 450 0.4044 0.6076 0.6076 0.4548 0.4548 0.5463 0.5463 0.2174 0.0 0.5 0.3931 nan

Framework versions

  • Transformers 4.16.2
  • Pytorch 1.10.2+cu113
  • Datasets 1.18.3
  • Tokenizers 0.11.0
Downloads last month
366
Hosted inference API
Text Classification
Examples
Examples
This model can be loaded on the Inference API on-demand.