Hugging Face's logo

tthhanh
/
legal_deberta

PyTorch
deberta-v2
Generated from Trainer
Model card Files Community
legal_deberta / README.md
tthhanh's picture
tthhanh
update model card README.md
ad4b9f3
preview code
|
raw
history blame
23.1 kB
metadata
license: mit
base_model: microsoft/deberta-v3-base
tags:
  - generated_from_trainer
model-index:
  - name: legal_deberta
    results: []

legal_deberta

This model is a fine-tuned version of microsoft/deberta-v3-base on an unknown dataset. It achieves the following results on the evaluation set:

  • Loss: 0.3745
  • Law Precision: 0.8118
  • Law Recall: 0.92
  • Law F1: 0.8625
  • Law Number: 75
  • Violated by Precision: 0.8434
  • Violated by Recall: 0.9333
  • Violated by F1: 0.8861
  • Violated by Number: 75
  • Violated on Precision: 0.7532
  • Violated on Recall: 0.7733
  • Violated on F1: 0.7632
  • Violated on Number: 75
  • Violation Precision: 0.5768
  • Violation Recall: 0.6705
  • Violation F1: 0.6201
  • Violation Number: 616
  • Overall Precision: 0.6348
  • Overall Recall: 0.7253
  • Overall F1: 0.6770
  • Overall Accuracy: 0.9495

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: 16
  • eval_batch_size: 16
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_steps: 500
  • num_epochs: 50

Training results

Training Loss Epoch Step Validation Loss Law Precision Law Recall Law F1 Law Number Violated by Precision Violated by Recall Violated by F1 Violated by Number Violated on Precision Violated on Recall Violated on F1 Violated on Number Violation Precision Violation Recall Violation F1 Violation Number Overall Precision Overall Recall Overall F1 Overall Accuracy
No log 1.0 45 0.8353 0.0 0.0 0.0 75 0.0 0.0 0.0 75 0.0 0.0 0.0 75 0.0 0.0 0.0 616 0.0 0.0 0.0 0.8060
No log 2.0 90 0.3620 0.0 0.0 0.0 75 0.0 0.0 0.0 75 0.0 0.0 0.0 75 0.1980 0.2873 0.2344 616 0.1928 0.2105 0.2013 0.8923
No log 3.0 135 0.2408 0.6582 0.6933 0.6753 75 0.6923 0.12 0.2045 75 0.0 0.0 0.0 75 0.3602 0.4432 0.3974 616 0.3920 0.3971 0.3946 0.9174
No log 4.0 180 0.2258 0.7229 0.8 0.7595 75 0.8788 0.7733 0.8227 75 0.3889 0.4667 0.4242 75 0.4976 0.5049 0.5012 616 0.5370 0.5517 0.5443 0.9346
No log 5.0 225 0.1876 0.6915 0.8667 0.7692 75 0.8406 0.7733 0.8056 75 0.4409 0.5467 0.4881 75 0.4716 0.5942 0.5259 616 0.5136 0.6302 0.5659 0.9439
No log 6.0 270 0.2125 0.6635 0.92 0.7709 75 0.7976 0.8933 0.8428 75 0.4842 0.6133 0.5412 75 0.4460 0.5763 0.5028 616 0.4977 0.6385 0.5594 0.9429
No log 7.0 315 0.2972 0.6091 0.8933 0.7243 75 0.7674 0.88 0.8199 75 0.5517 0.64 0.5926 75 0.4647 0.5455 0.5019 616 0.5139 0.6147 0.5598 0.9397
No log 8.0 360 0.2361 0.8333 0.8667 0.8497 75 0.8831 0.9067 0.8947 75 0.5889 0.7067 0.6424 75 0.5013 0.6169 0.5531 616 0.5643 0.6730 0.6139 0.9472
No log 9.0 405 0.2956 0.5156 0.88 0.6502 75 0.7263 0.92 0.8118 75 0.5096 0.7067 0.5922 75 0.5086 0.5731 0.5389 616 0.5299 0.6433 0.5811 0.9391
No log 10.0 450 0.2722 0.7097 0.88 0.7857 75 0.9079 0.92 0.9139 75 0.6049 0.6533 0.6282 75 0.5157 0.6396 0.5710 616 0.5700 0.6873 0.6232 0.9431
No log 11.0 495 0.3221 0.7234 0.9067 0.8047 75 0.8919 0.88 0.8859 75 0.7105 0.72 0.7152 75 0.5030 0.5422 0.5219 616 0.5749 0.6207 0.5969 0.9374
0.271 12.0 540 0.3475 0.6538 0.9067 0.7598 75 0.6542 0.9333 0.7692 75 0.4831 0.76 0.5907 75 0.5141 0.5341 0.5239 616 0.5408 0.6231 0.5790 0.9344
0.271 13.0 585 0.2667 0.7128 0.8933 0.7929 75 0.6 0.92 0.7263 75 0.5824 0.7067 0.6386 75 0.5530 0.6526 0.5987 616 0.5755 0.7027 0.6328 0.9484
0.271 14.0 630 0.2868 0.7113 0.92 0.8023 75 0.8642 0.9333 0.8974 75 0.6706 0.76 0.7125 75 0.5393 0.6347 0.5831 616 0.5941 0.6980 0.6419 0.9508
0.271 15.0 675 0.3035 0.7419 0.92 0.8214 75 0.8537 0.9333 0.8917 75 0.6471 0.7333 0.6875 75 0.5681 0.6769 0.6178 616 0.6147 0.7265 0.6659 0.9475
0.271 16.0 720 0.2746 0.7053 0.8933 0.7882 75 0.8861 0.9333 0.9091 75 0.6875 0.7333 0.7097 75 0.5654 0.6737 0.6148 616 0.6144 0.7218 0.6638 0.9523
0.271 17.0 765 0.2846 0.7907 0.9067 0.8447 75 0.9221 0.9467 0.9342 75 0.6709 0.7067 0.6883 75 0.5904 0.6623 0.6243 616 0.6431 0.7134 0.6764 0.9506
0.271 18.0 810 0.3004 0.75 0.92 0.8263 75 0.8434 0.9333 0.8861 75 0.6747 0.7467 0.7089 75 0.5501 0.6769 0.6070 616 0.6024 0.7277 0.6591 0.9464
0.271 19.0 855 0.3235 0.6832 0.92 0.7841 75 0.8214 0.92 0.8679 75 0.6667 0.7733 0.7160 75 0.5451 0.6867 0.6078 616 0.5906 0.7360 0.6554 0.9482
0.271 20.0 900 0.3274 0.7391 0.9067 0.8144 75 0.8861 0.9333 0.9091 75 0.6835 0.72 0.7013 75 0.5679 0.6786 0.6183 616 0.6187 0.7253 0.6678 0.9474
0.271 21.0 945 0.3756 0.7882 0.8933 0.8375 75 0.8537 0.9333 0.8917 75 0.6962 0.7333 0.7143 75 0.6046 0.6429 0.6231 616 0.6526 0.6992 0.6751 0.9468
0.271 22.0 990 0.3511 0.8072 0.8933 0.8481 75 0.8974 0.9333 0.9150 75 0.7215 0.76 0.7403 75 0.5813 0.6445 0.6112 616 0.6403 0.7027 0.6701 0.9460
0.007 23.0 1035 0.3187 0.8375 0.8933 0.8645 75 0.9079 0.92 0.9139 75 0.7237 0.7333 0.7285 75 0.5816 0.6656 0.6207 616 0.6414 0.7146 0.6760 0.9507
0.007 24.0 1080 0.3383 0.8171 0.8933 0.8535 75 0.8118 0.92 0.8625 75 0.6747 0.7467 0.7089 75 0.5727 0.6331 0.6014 616 0.6251 0.6920 0.6569 0.9473
0.007 25.0 1125 0.3231 0.7952 0.88 0.8354 75 0.8434 0.9333 0.8861 75 0.6667 0.72 0.6923 75 0.5524 0.6851 0.6116 616 0.6053 0.7277 0.6609 0.9483
0.007 26.0 1170 0.3099 0.7033 0.8533 0.7711 75 0.8642 0.9333 0.8974 75 0.7143 0.7333 0.7237 75 0.5722 0.6753 0.6195 616 0.6199 0.7194 0.6659 0.9509
0.007 27.0 1215 0.3202 0.7701 0.8933 0.8272 75 0.8961 0.92 0.9079 75 0.6582 0.6933 0.6753 75 0.5977 0.6705 0.6320 616 0.6435 0.7146 0.6772 0.9509
0.007 28.0 1260 0.3381 0.7263 0.92 0.8118 75 0.7931 0.92 0.8519 75 0.7 0.7467 0.7226 75 0.5909 0.6753 0.6303 616 0.6315 0.7253 0.6752 0.9496
0.007 29.0 1305 0.3413 0.7841 0.92 0.8466 75 0.7955 0.9333 0.8589 75 0.6463 0.7067 0.6752 75 0.5895 0.6737 0.6288 616 0.6310 0.7218 0.6733 0.95
0.007 30.0 1350 0.3427 0.8023 0.92 0.8571 75 0.8961 0.92 0.9079 75 0.6962 0.7333 0.7143 75 0.5974 0.6721 0.6325 616 0.6492 0.7218 0.6836 0.9508
0.007 31.0 1395 0.3473 0.8023 0.92 0.8571 75 0.8861 0.9333 0.9091 75 0.7143 0.7333 0.7237 75 0.5738 0.6623 0.6149 616 0.6317 0.7158 0.6711 0.9493
0.007 32.0 1440 0.3531 0.7188 0.92 0.8070 75 0.8140 0.9333 0.8696 75 0.7195 0.7867 0.7516 75 0.5740 0.6672 0.6171 616 0.6214 0.7241 0.6689 0.9495
0.007 33.0 1485 0.3556 0.7263 0.92 0.8118 75 0.8140 0.9333 0.8696 75 0.7108 0.7867 0.7468 75 0.5783 0.6656 0.6189 616 0.6249 0.7229 0.6703 0.9492
0.0009 34.0 1530 0.3569 0.7667 0.92 0.8364 75 0.8140 0.9333 0.8696 75 0.7215 0.76 0.7403 75 0.5783 0.6656 0.6189 616 0.6286 0.7206 0.6715 0.9493
0.0009 35.0 1575 0.3630 0.7841 0.92 0.8466 75 0.8140 0.9333 0.8696 75 0.7 0.7467 0.7226 75 0.5838 0.6672 0.6227 616 0.6326 0.7206 0.6737 0.9489
0.0009 36.0 1620 0.3624 0.7667 0.92 0.8364 75 0.8140 0.9333 0.8696 75 0.7179 0.7467 0.7320 75 0.5782 0.6721 0.6216 616 0.6278 0.7241 0.6726 0.9493
0.0009 37.0 1665 0.3614 0.7667 0.92 0.8364 75 0.8140 0.9333 0.8696 75 0.7403 0.76 0.75 75 0.5744 0.6705 0.6187 616 0.6265 0.7241 0.6718 0.9493
0.0009 38.0 1710 0.3630 0.7841 0.92 0.8466 75 0.8235 0.9333 0.8750 75 0.7662 0.7867 0.7763 75 0.5842 0.6705 0.6243 616 0.6385 0.7265 0.6796 0.9496
0.0009 39.0 1755 0.3645 0.8118 0.92 0.8625 75 0.8235 0.9333 0.8750 75 0.7662 0.7867 0.7763 75 0.5842 0.6705 0.6243 616 0.6405 0.7265 0.6808 0.9499
0.0009 40.0 1800 0.3670 0.8313 0.92 0.8734 75 0.8333 0.9333 0.8805 75 0.7632 0.7733 0.7682 75 0.5784 0.6705 0.6211 616 0.6374 0.7253 0.6785 0.9499
0.0009 41.0 1845 0.3693 0.8214 0.92 0.8679 75 0.8333 0.9333 0.8805 75 0.7632 0.7733 0.7682 75 0.5831 0.6721 0.6244 616 0.6405 0.7265 0.6808 0.9497
0.0009 42.0 1890 0.3727 0.7931 0.92 0.8519 75 0.8235 0.9333 0.8750 75 0.7436 0.7733 0.7582 75 0.5946 0.6737 0.6317 616 0.6456 0.7277 0.6842 0.9491
0.0009 43.0 1935 0.3739 0.8023 0.92 0.8571 75 0.8235 0.9333 0.8750 75 0.7160 0.7733 0.7436 75 0.5968 0.6753 0.6337 616 0.6459 0.7289 0.6849 0.9490
0.0009 44.0 1980 0.3750 0.8023 0.92 0.8571 75 0.8235 0.9333 0.8750 75 0.7 0.7467 0.7226 75 0.5968 0.6753 0.6337 616 0.6445 0.7265 0.6831 0.9490
0.0001 45.0 2025 0.3753 0.8023 0.92 0.8571 75 0.8235 0.9333 0.8750 75 0.7160 0.7733 0.7436 75 0.5932 0.6769 0.6323 616 0.6429 0.7301 0.6837 0.9490
0.0001 46.0 2070 0.3755 0.8023 0.92 0.8571 75 0.8235 0.9333 0.8750 75 0.7342 0.7733 0.7532 75 0.5912 0.6786 0.6319 616 0.6426 0.7313 0.6841 0.9492
0.0001 47.0 2115 0.3759 0.8023 0.92 0.8571 75 0.8235 0.9333 0.8750 75 0.7342 0.7733 0.7532 75 0.5896 0.6786 0.6309 616 0.6413 0.7313 0.6833 0.9492
0.0001 48.0 2160 0.3742 0.8118 0.92 0.8625 75 0.8434 0.9333 0.8861 75 0.7532 0.7733 0.7632 75 0.5770 0.6688 0.6195 616 0.6350 0.7241 0.6767 0.9493
0.0001 49.0 2205 0.3744 0.8118 0.92 0.8625 75 0.8434 0.9333 0.8861 75 0.7532 0.7733 0.7632 75 0.5760 0.6705 0.6197 616 0.6341 0.7253 0.6767 0.9495
0.0001 50.0 2250 0.3745 0.8118 0.92 0.8625 75 0.8434 0.9333 0.8861 75 0.7532 0.7733 0.7632 75 0.5768 0.6705 0.6201 616 0.6348 0.7253 0.6770 0.9495

Framework versions

  • Transformers 4.31.0
  • Pytorch 2.4.0+cu121
  • Datasets 2.15.0
  • Tokenizers 0.13.3