ananth-docai2 / README.md
ananthrgv's picture
End of training
d881c86
metadata
license: mit
tags:
  - generated_from_trainer
datasets:
  - funsd-layoutlmv3
model-index:
  - name: ananth-docai2
    results: []

ananth-docai2

This model is a fine-tuned version of SCUT-DLVCLab/lilt-roberta-en-base on the funsd-layoutlmv3 dataset. It achieves the following results on the evaluation set:

  • Loss: 1.4203
  • Answer: {'precision': 0.8505747126436781, 'recall': 0.9057527539779682, 'f1': 0.8772969768820391, 'number': 817}
  • Header: {'precision': 0.6476190476190476, 'recall': 0.5714285714285714, 'f1': 0.6071428571428571, 'number': 119}
  • Question: {'precision': 0.9104477611940298, 'recall': 0.9062209842154132, 'f1': 0.9083294555607259, 'number': 1077}
  • Overall Precision: 0.8715
  • Overall Recall: 0.8862
  • Overall F1: 0.8788
  • Overall Accuracy: 0.8269

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
  • training_steps: 2500
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Answer Header Question Overall Precision Overall Recall Overall F1 Overall Accuracy
0.4218 10.53 200 1.0024 {'precision': 0.8727272727272727, 'recall': 0.8812729498164015, 'f1': 0.8769792935444579, 'number': 817} {'precision': 0.4036144578313253, 'recall': 0.5630252100840336, 'f1': 0.47017543859649125, 'number': 119} {'precision': 0.8674812030075187, 'recall': 0.8570102135561746, 'f1': 0.8622139187295657, 'number': 1077} 0.8321 0.8495 0.8407 0.7973
0.0532 21.05 400 1.1791 {'precision': 0.8563218390804598, 'recall': 0.9118727050183598, 'f1': 0.8832246591582691, 'number': 817} {'precision': 0.5486725663716814, 'recall': 0.5210084033613446, 'f1': 0.5344827586206897, 'number': 119} {'precision': 0.9044943820224719, 'recall': 0.8969359331476323, 'f1': 0.9006993006993008, 'number': 1077} 0.8645 0.8808 0.8725 0.8103
0.0117 31.58 600 1.5177 {'precision': 0.8064516129032258, 'recall': 0.9179926560587516, 'f1': 0.8586147681740126, 'number': 817} {'precision': 0.6046511627906976, 'recall': 0.4369747899159664, 'f1': 0.5073170731707317, 'number': 119} {'precision': 0.9019607843137255, 'recall': 0.8542246982358404, 'f1': 0.8774439675727229, 'number': 1077} 0.8458 0.8554 0.8506 0.7952
0.0067 42.11 800 1.4884 {'precision': 0.8443935926773455, 'recall': 0.9033047735618115, 'f1': 0.872856298048492, 'number': 817} {'precision': 0.515625, 'recall': 0.5546218487394958, 'f1': 0.5344129554655871, 'number': 119} {'precision': 0.8784530386740331, 'recall': 0.8857938718662952, 'f1': 0.8821081830790567, 'number': 1077} 0.8420 0.8733 0.8574 0.7963
0.0034 52.63 1000 1.4203 {'precision': 0.8505747126436781, 'recall': 0.9057527539779682, 'f1': 0.8772969768820391, 'number': 817} {'precision': 0.6476190476190476, 'recall': 0.5714285714285714, 'f1': 0.6071428571428571, 'number': 119} {'precision': 0.9104477611940298, 'recall': 0.9062209842154132, 'f1': 0.9083294555607259, 'number': 1077} 0.8715 0.8862 0.8788 0.8269
0.0023 63.16 1200 1.5225 {'precision': 0.834096109839817, 'recall': 0.8922888616891065, 'f1': 0.8622117090479007, 'number': 817} {'precision': 0.5689655172413793, 'recall': 0.5546218487394958, 'f1': 0.5617021276595745, 'number': 119} {'precision': 0.8962001853568119, 'recall': 0.8978644382544104, 'f1': 0.8970315398886828, 'number': 1077} 0.8516 0.8753 0.8633 0.8096
0.0013 73.68 1400 1.6801 {'precision': 0.848, 'recall': 0.9082007343941249, 'f1': 0.8770685579196217, 'number': 817} {'precision': 0.6741573033707865, 'recall': 0.5042016806722689, 'f1': 0.576923076923077, 'number': 119} {'precision': 0.8977695167286245, 'recall': 0.8969359331476323, 'f1': 0.8973525313516025, 'number': 1077} 0.8667 0.8783 0.8724 0.7977
0.0014 84.21 1600 1.6236 {'precision': 0.8876543209876543, 'recall': 0.8800489596083231, 'f1': 0.8838352796558081, 'number': 817} {'precision': 0.6237623762376238, 'recall': 0.5294117647058824, 'f1': 0.5727272727272728, 'number': 119} {'precision': 0.8656330749354005, 'recall': 0.9331476323119777, 'f1': 0.8981233243967828, 'number': 1077} 0.8625 0.8877 0.8749 0.8072
0.0006 94.74 1800 1.7231 {'precision': 0.8619883040935673, 'recall': 0.9020807833537332, 'f1': 0.881578947368421, 'number': 817} {'precision': 0.6883116883116883, 'recall': 0.44537815126050423, 'f1': 0.5408163265306123, 'number': 119} {'precision': 0.8748890860692103, 'recall': 0.9155060352831941, 'f1': 0.8947368421052633, 'number': 1077} 0.8626 0.8823 0.8723 0.8019
0.0005 105.26 2000 1.8217 {'precision': 0.8342665173572228, 'recall': 0.9118727050183598, 'f1': 0.871345029239766, 'number': 817} {'precision': 0.6, 'recall': 0.5042016806722689, 'f1': 0.547945205479452, 'number': 119} {'precision': 0.9049858889934148, 'recall': 0.89322191272052, 'f1': 0.8990654205607476, 'number': 1077} 0.8594 0.8778 0.8685 0.7964
0.0004 115.79 2200 1.7688 {'precision': 0.8561484918793504, 'recall': 0.9033047735618115, 'f1': 0.8790946992257296, 'number': 817} {'precision': 0.6555555555555556, 'recall': 0.4957983193277311, 'f1': 0.5645933014354068, 'number': 119} {'precision': 0.8827272727272727, 'recall': 0.9015784586815228, 'f1': 0.8920532843362425, 'number': 1077} 0.8616 0.8783 0.8699 0.7956
0.0002 126.32 2400 1.7726 {'precision': 0.8458904109589042, 'recall': 0.9069767441860465, 'f1': 0.8753691671588896, 'number': 817} {'precision': 0.6741573033707865, 'recall': 0.5042016806722689, 'f1': 0.576923076923077, 'number': 119} {'precision': 0.8878676470588235, 'recall': 0.8969359331476323, 'f1': 0.892378752886836, 'number': 1077} 0.8607 0.8778 0.8692 0.7961

Framework versions

  • Transformers 4.25.1
  • Pytorch 1.13.1+cu117
  • Datasets 2.7.1
  • Tokenizers 0.13.2