henryscheible's picture
update model card README.md
179fc25
metadata
license: apache-2.0
tags:
  - generated_from_trainer
datasets:
  - crows_pairs
metrics:
  - accuracy
model-index:
  - name: bert-base-uncased_crows_pairs_classifieronly
    results:
      - task:
          name: Text Classification
          type: text-classification
        dataset:
          name: crows_pairs
          type: crows_pairs
          config: crows_pairs
          split: test
          args: crows_pairs
        metrics:
          - name: Accuracy
            type: accuracy
            value: 0.5364238410596026

bert-base-uncased_crows_pairs_classifieronly

This model is a fine-tuned version of bert-base-uncased on the crows_pairs dataset. It achieves the following results on the evaluation set:

  • Loss: 0.6924
  • Accuracy: 0.5364
  • Tp: 0.0066
  • Tn: 0.5298
  • Fp: 0.0033
  • Fn: 0.4603

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: 0.0001
  • train_batch_size: 64
  • eval_batch_size: 64
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 50

Training results

Training Loss Epoch Step Validation Loss Accuracy Tp Tn Fp Fn
0.7219 1.05 20 0.7045 0.5331 0.0 0.5331 0.0 0.4669
0.6962 2.11 40 0.6962 0.4669 0.4669 0.0 0.5331 0.0
0.6983 3.16 60 0.6925 0.5331 0.0 0.5331 0.0 0.4669
0.7018 4.21 80 0.6962 0.4669 0.4669 0.0 0.5331 0.0
0.6972 5.26 100 0.6915 0.5331 0.0 0.5331 0.0 0.4669
0.6964 6.32 120 0.6934 0.4801 0.1391 0.3411 0.1921 0.3278
0.6983 7.37 140 0.6940 0.4636 0.3709 0.0927 0.4404 0.0960
0.7025 8.42 160 0.6964 0.4669 0.4669 0.0 0.5331 0.0
0.6958 9.47 180 0.6919 0.5331 0.0 0.5331 0.0 0.4669
0.7079 10.53 200 0.7002 0.4669 0.4669 0.0 0.5331 0.0
0.7033 11.58 220 0.6915 0.5331 0.0 0.5331 0.0 0.4669
0.6932 12.63 240 0.6933 0.5 0.1060 0.3940 0.1391 0.3609
0.7075 13.68 260 0.6919 0.5331 0.0 0.5331 0.0 0.4669
0.695 14.74 280 0.6936 0.4371 0.1523 0.2848 0.2483 0.3146
0.7068 15.79 300 0.6916 0.5331 0.0 0.5331 0.0 0.4669
0.7007 16.84 320 0.6916 0.5331 0.0 0.5331 0.0 0.4669
0.7035 17.89 340 0.6961 0.4669 0.4669 0.0 0.5331 0.0
0.7002 18.95 360 0.6919 0.5331 0.0 0.5331 0.0 0.4669
0.6992 20.0 380 0.6930 0.5166 0.0331 0.4834 0.0497 0.4338
0.7024 21.05 400 0.6924 0.5364 0.0066 0.5298 0.0033 0.4603
0.694 22.11 420 0.6949 0.4669 0.4603 0.0066 0.5265 0.0066
0.7085 23.16 440 0.6928 0.5265 0.0199 0.5066 0.0265 0.4470
0.6999 24.21 460 0.6936 0.4338 0.1457 0.2881 0.2450 0.3212
0.6926 25.26 480 0.6921 0.5331 0.0033 0.5298 0.0033 0.4636
0.7088 26.32 500 0.6923 0.5331 0.0033 0.5298 0.0033 0.4636
0.6932 27.37 520 0.6922 0.5331 0.0033 0.5298 0.0033 0.4636
0.7011 28.42 540 0.6925 0.5364 0.0066 0.5298 0.0033 0.4603
0.7016 29.47 560 0.6923 0.5331 0.0033 0.5298 0.0033 0.4636
0.7015 30.53 580 0.6925 0.5364 0.0099 0.5265 0.0066 0.4570
0.7002 31.58 600 0.6929 0.5232 0.0331 0.4901 0.0430 0.4338
0.701 32.63 620 0.6932 0.5099 0.0563 0.4536 0.0795 0.4106
0.693 33.68 640 0.6921 0.5331 0.0 0.5331 0.0 0.4669
0.711 34.74 660 0.6925 0.5364 0.0099 0.5265 0.0066 0.4570
0.7013 35.79 680 0.6924 0.5331 0.0033 0.5298 0.0033 0.4636
0.6975 36.84 700 0.6916 0.5331 0.0 0.5331 0.0 0.4669
0.7035 37.89 720 0.6918 0.5331 0.0 0.5331 0.0 0.4669
0.6991 38.95 740 0.6929 0.5232 0.0298 0.4934 0.0397 0.4371
0.7165 40.0 760 0.6923 0.5331 0.0033 0.5298 0.0033 0.4636
0.7029 41.05 780 0.6931 0.5066 0.0464 0.4603 0.0728 0.4205
0.7021 42.11 800 0.6923 0.5331 0.0033 0.5298 0.0033 0.4636
0.6993 43.16 820 0.6935 0.4934 0.1291 0.3642 0.1689 0.3377
0.7 44.21 840 0.6926 0.5331 0.0132 0.5199 0.0132 0.4536
0.7023 45.26 860 0.6926 0.5331 0.0099 0.5232 0.0099 0.4570
0.6961 46.32 880 0.6927 0.5232 0.0132 0.5099 0.0232 0.4536
0.7014 47.37 900 0.6923 0.5331 0.0033 0.5298 0.0033 0.4636
0.7025 48.42 920 0.6924 0.5331 0.0033 0.5298 0.0033 0.4636
0.702 49.47 940 0.6924 0.5364 0.0066 0.5298 0.0033 0.4603

Framework versions

  • Transformers 4.26.1
  • Pytorch 1.13.1
  • Datasets 2.10.1
  • Tokenizers 0.13.2