Edit model card

Tagset

  • O
  • B-CITATION
  • I-CITATION
  • B-LAW
  • I-LAW

Training

  • The model was trained with the following hyperparamters:
    • batch size: 64
    • learning_rate: 0.00001
    • number of training epochs: 50 (actually trained: 23)
    • early stopping patience: 5

Predict scores

metric score
de_predict/_CITATION_f1 97.93
de_predict/_CITATION_precision 98.53
de_predict/_CITATION_recall 97.34
de_predict/_LAW_f1 92.08
de_predict/_LAW_precision 85.99
de_predict/_LAW_recall 99.1
de_predict/_accuracy_normalized 98.8
de_predict/_macro-f1 95.04
de_predict/_macro-precision 98.22
de_predict/_macro-recall 92.32
de_predict/_micro-f1 94.06
de_predict/_micro-precision 98.49
de_predict/_micro-recall 90.01
de_predict/_steps_per_second 54.9
de_predict/_weighted-f1 93.97
de_predict/_weighted-precision 98.55
de_predict/_weighted-recall 90.01
fr_predict/_CITATION_f1 95.55
fr_predict/_CITATION_precision 96.85
fr_predict/_CITATION_recall 94.28
fr_predict/_LAW_f1 91.01
fr_predict/_LAW_precision 83.67
fr_predict/_LAW_recall 99.76
fr_predict/_accuracy_normalized 98.31
fr_predict/_macro-f1 93.3
fr_predict/_macro-precision 97.02
fr_predict/_macro-recall 90.3
fr_predict/_micro-f1 92.06
fr_predict/_micro-precision 98.42
fr_predict/_micro-recall 86.47
fr_predict/_steps_per_second 59.3
fr_predict/_weighted-f1 91.99
fr_predict/_weighted-precision 98.62
fr_predict/_weighted-recall 86.47
it_predict/_CITATION_f1 97.04
it_predict/_CITATION_precision 97.7
it_predict/_CITATION_recall 96.39
it_predict/_LAW_f1 90.99
it_predict/_LAW_precision 84.23
it_predict/_LAW_recall 98.94
it_predict/_accuracy_normalized 98.92
it_predict/_macro-f1 94.13
it_predict/_macro-precision 97.66
it_predict/_macro-recall 91.2
it_predict/_micro-f1 93.11
it_predict/_micro-precision 98.03
it_predict/_micro-recall 88.67
it_predict/_steps_per_second 56.3
it_predict/_weighted-f1 93
it_predict/_weighted-precision 98.13
it_predict/_weighted-recall 88.67
predict/_CITATION_f1 97.36
predict/_CITATION_precision 98.11
predict/_CITATION_recall 96.62
predict/_LAW_f1 91.68
predict/_LAW_precision 85.15
predict/_LAW_recall 99.3
predict/_accuracy_normalized 98.68
predict/_macro-f1 94.56
predict/_macro-precision 97.96
predict/_macro-recall 91.7
predict/_micro-f1 93.43
predict/_micro-precision 98.45
predict/_micro-recall 88.91
predict/_steps_per_second 55.7
predict/_weighted-f1 93.34
predict/_weighted-precision 98.54
predict/_weighted-recall 88.91
predict_samples 28218
Downloads last month
11
Safetensors
Model size
118M params
Tensor type
I64
·
FP16
·