lmv2-g-rai2-732-doc-10-07

This model is a fine-tuned version of microsoft/layoutlmv2-base-uncased on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 0.0222
  • Dob Key Precision: 0.7810
  • Dob Key Recall: 0.8359
  • Dob Key F1: 0.8075
  • Dob Key Number: 128
  • Dob Value Precision: 0.9767
  • Dob Value Recall: 0.9618
  • Dob Value F1: 0.9692
  • Dob Value Number: 131
  • Doctor Name Key Precision: 0.5714
  • Doctor Name Key Recall: 0.6197
  • Doctor Name Key F1: 0.5946
  • Doctor Name Key Number: 71
  • Doctor Name Value Precision: 0.8861
  • Doctor Name Value Recall: 0.9091
  • Doctor Name Value F1: 0.8974
  • Doctor Name Value Number: 77
  • Patient Name Key Precision: 0.7639
  • Patient Name Key Recall: 0.7639
  • Patient Name Key F1: 0.7639
  • Patient Name Key Number: 144
  • Patient Name Value Precision: 0.9595
  • Patient Name Value Recall: 0.9530
  • Patient Name Value F1: 0.9562
  • Patient Name Value Number: 149
  • Overall Precision: 0.8389
  • Overall Recall: 0.8557
  • Overall F1: 0.8472
  • Overall Accuracy: 0.9939

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: 4e-05
  • train_batch_size: 1
  • eval_batch_size: 1
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: constant
  • num_epochs: 30

Training results

Training Loss Epoch Step Validation Loss Dob Key Precision Dob Key Recall Dob Key F1 Dob Key Number Dob Value Precision Dob Value Recall Dob Value F1 Dob Value Number Doctor Name Key Precision Doctor Name Key Recall Doctor Name Key F1 Doctor Name Key Number Doctor Name Value Precision Doctor Name Value Recall Doctor Name Value F1 Doctor Name Value Number Patient Name Key Precision Patient Name Key Recall Patient Name Key F1 Patient Name Key Number Patient Name Value Precision Patient Name Value Recall Patient Name Value F1 Patient Name Value Number Overall Precision Overall Recall Overall F1 Overall Accuracy
0.6913 1.0 585 0.1911 0.0 0.0 0.0 128 0.0 0.0 0.0 131 0.0 0.0 0.0 71 0.0 0.0 0.0 77 0.0 0.0 0.0 144 0.0 0.0 0.0 149 0.0 0.0 0.0 0.9671
0.1083 2.0 1170 0.0588 0.7692 0.7812 0.7752 128 0.9615 0.9542 0.9579 131 0.5645 0.4930 0.5263 71 0.9178 0.8701 0.8933 77 0.7397 0.75 0.7448 144 0.9346 0.9597 0.9470 149 0.8329 0.8257 0.8293 0.9936
0.0515 3.0 1755 0.0377 0.7803 0.8047 0.7923 128 0.875 0.9084 0.8914 131 0.5867 0.6197 0.6027 71 0.9231 0.9351 0.9290 77 0.6903 0.7431 0.7157 144 0.9732 0.9732 0.9732 149 0.8138 0.8429 0.8281 0.9942
0.0348 4.0 2340 0.0287 0.7710 0.7891 0.7799 128 0.9545 0.9618 0.9582 131 0.5789 0.6197 0.5986 71 0.8391 0.9481 0.8902 77 0.75 0.75 0.75 144 0.9603 0.9732 0.9667 149 0.8280 0.8529 0.8403 0.9940
0.0279 5.0 2925 0.0262 0.6541 0.8125 0.7247 128 0.9466 0.9466 0.9466 131 0.6027 0.6197 0.6111 71 0.6915 0.8442 0.7602 77 0.7397 0.75 0.7448 144 0.9062 0.9732 0.9385 149 0.7733 0.8429 0.8066 0.9938
0.0232 6.0 3510 0.0263 0.6603 0.8047 0.7254 128 0.9767 0.9618 0.9692 131 0.6143 0.6056 0.6099 71 0.9315 0.8831 0.9067 77 0.75 0.75 0.75 144 0.9728 0.9597 0.9662 149 0.8220 0.8443 0.8330 0.9941
0.0204 7.0 4095 0.0220 0.8062 0.8125 0.8093 128 0.9767 0.9618 0.9692 131 0.5946 0.6197 0.6069 71 0.8590 0.8701 0.8645 77 0.75 0.75 0.75 144 0.9603 0.9732 0.9667 149 0.8426 0.8486 0.8456 0.9945
0.0202 8.0 4680 0.0217 0.7591 0.8125 0.7849 128 0.9769 0.9695 0.9732 131 0.6027 0.6197 0.6111 71 0.8625 0.8961 0.8790 77 0.5989 0.7361 0.6604 144 0.9664 0.9664 0.9664 149 0.7962 0.8486 0.8216 0.9939
0.0174 9.0 5265 0.0219 0.8062 0.8125 0.8093 128 0.9695 0.9695 0.9695 131 0.5714 0.6197 0.5946 71 0.8452 0.9221 0.8820 77 0.75 0.75 0.75 144 0.9177 0.9732 0.9446 149 0.8285 0.8557 0.8419 0.9938
0.0172 10.0 5850 0.0209 0.7803 0.8047 0.7923 128 0.9695 0.9695 0.9695 131 0.5789 0.6197 0.5986 71 0.8675 0.9351 0.9 77 0.75 0.75 0.75 144 0.9667 0.9732 0.9699 149 0.8366 0.8557 0.8460 0.9941
0.0174 11.0 6435 0.0206 0.8062 0.8125 0.8093 128 0.9767 0.9618 0.9692 131 0.3554 0.6056 0.4479 71 0.8889 0.9351 0.9114 77 0.75 0.75 0.75 144 0.9664 0.9664 0.9664 149 0.7928 0.8529 0.8217 0.9944
0.0156 12.0 7020 0.0219 0.7132 0.7578 0.7348 128 0.9767 0.9618 0.9692 131 0.4884 0.5915 0.5350 71 0.9114 0.9351 0.9231 77 0.5856 0.7361 0.6523 144 0.9732 0.9732 0.9732 149 0.7737 0.84 0.8055 0.9936
0.0158 13.0 7605 0.0203 0.6190 0.8125 0.7027 128 0.9767 0.9618 0.9692 131 0.6027 0.6197 0.6111 71 0.9114 0.9351 0.9231 77 0.6044 0.7639 0.6748 144 0.96 0.9664 0.9632 149 0.7682 0.8571 0.8103 0.9940
0.0152 14.0 8190 0.0195 0.7252 0.7422 0.7336 128 0.9767 0.9618 0.9692 131 0.6027 0.6197 0.6111 71 0.8765 0.9221 0.8987 77 0.75 0.75 0.75 144 0.9664 0.9664 0.9664 149 0.8317 0.84 0.8358 0.9940
0.0158 15.0 8775 0.0204 0.7252 0.7422 0.7336 128 0.9767 0.9618 0.9692 131 0.5658 0.6056 0.5850 71 0.8861 0.9091 0.8974 77 0.6424 0.7361 0.6861 144 0.9664 0.9664 0.9664 149 0.8011 0.8343 0.8174 0.9939
0.0172 16.0 9360 0.0218 0.5741 0.7266 0.6414 128 0.9767 0.9618 0.9692 131 0.5676 0.5915 0.5793 71 0.8987 0.9221 0.9103 77 0.6145 0.7639 0.6811 144 0.9664 0.9664 0.9664 149 0.7591 0.8371 0.7962 0.9929
0.0137 17.0 9945 0.0220 0.8 0.8125 0.8062 128 0.9767 0.9618 0.9692 131 0.5946 0.6197 0.6069 71 0.8734 0.8961 0.8846 77 0.6485 0.7431 0.6926 144 0.9536 0.9664 0.9600 149 0.8159 0.8486 0.8319 0.9942
0.0143 18.0 10530 0.0222 0.7810 0.8359 0.8075 128 0.9767 0.9618 0.9692 131 0.5714 0.6197 0.5946 71 0.8861 0.9091 0.8974 77 0.7639 0.7639 0.7639 144 0.9595 0.9530 0.9562 149 0.8389 0.8557 0.8472 0.9939
0.0137 19.0 11115 0.0222 0.7557 0.7734 0.7645 128 0.9767 0.9618 0.9692 131 0.5395 0.5775 0.5578 71 0.8608 0.8831 0.8718 77 0.6732 0.7153 0.6936 144 0.96 0.9664 0.9632 149 0.8092 0.83 0.8195 0.9935
0.0126 20.0 11700 0.0234 0.7907 0.7969 0.7938 128 0.9612 0.9466 0.9538 131 0.5309 0.6056 0.5658 71 0.8372 0.9351 0.8834 77 0.5917 0.6944 0.6390 144 0.96 0.9664 0.9632 149 0.7863 0.8357 0.8102 0.9933
0.0122 21.0 12285 0.0228 0.6350 0.6797 0.6566 128 0.9767 0.9618 0.9692 131 0.5 0.6197 0.5535 71 0.8659 0.9221 0.8931 77 0.6824 0.7014 0.6918 144 0.9536 0.9664 0.9600 149 0.7796 0.8186 0.7986 0.9930
0.0114 22.0 12870 0.0230 0.7863 0.8047 0.7954 128 0.9615 0.9542 0.9579 131 0.38 0.5352 0.4444 71 0.875 0.9091 0.8917 77 0.6689 0.7014 0.6847 144 0.9533 0.9597 0.9565 149 0.7817 0.8286 0.8044 0.9938
0.0112 23.0 13455 0.0259 0.5038 0.5156 0.5097 128 0.9690 0.9542 0.9615 131 0.5811 0.6056 0.5931 71 0.8987 0.9221 0.9103 77 0.7361 0.7361 0.7361 144 0.9664 0.9664 0.9664 149 0.7861 0.7929 0.7895 0.9925
0.0108 24.0 14040 0.0280 0.6591 0.6797 0.6692 128 0.9690 0.9542 0.9615 131 0.4681 0.6197 0.5333 71 0.8481 0.8701 0.8590 77 0.7241 0.7292 0.7266 144 0.94 0.9463 0.9431 149 0.7805 0.8129 0.7964 0.9929
0.0085 25.0 14625 0.0277 0.6618 0.7031 0.6818 128 0.9615 0.9542 0.9579 131 0.4607 0.5775 0.5125 71 0.8642 0.9091 0.8861 77 0.6733 0.7014 0.6871 144 0.96 0.9664 0.9632 149 0.7758 0.8157 0.7953 0.9932
0.0081 26.0 15210 0.0255 0.7634 0.7812 0.7722 128 0.9690 0.9542 0.9615 131 0.5783 0.6761 0.6234 71 0.9125 0.9481 0.9299 77 0.6918 0.7014 0.6966 144 0.9470 0.9597 0.9533 149 0.8194 0.8429 0.8310 0.9940
0.0063 27.0 15795 0.0302 0.6977 0.7031 0.7004 128 0.9767 0.9618 0.9692 131 0.5325 0.5775 0.5541 71 0.8519 0.8961 0.8734 77 0.7222 0.7222 0.7222 144 0.9470 0.9597 0.9533 149 0.8059 0.8186 0.8122 0.9933
0.005 28.0 16380 0.0306 0.7829 0.7891 0.7860 128 0.9767 0.9618 0.9692 131 0.5679 0.6479 0.6053 71 0.8193 0.8831 0.85 77 0.7172 0.7222 0.7197 144 0.96 0.9664 0.9632 149 0.8215 0.8414 0.8313 0.9942
0.0158 29.0 16965 0.2296 0.8 0.0312 0.0602 128 1.0 0.0153 0.0301 131 0.0 0.0 0.0 71 0.0 0.0 0.0 77 0.1724 0.0694 0.0990 144 0.0616 0.1477 0.0870 149 0.0862 0.0543 0.0666 0.9655
0.1003 30.0 17550 0.0310 0.7385 0.75 0.7442 128 0.9542 0.9542 0.9542 131 0.6184 0.6620 0.6395 71 0.8675 0.9351 0.9 77 0.75 0.75 0.75 144 0.9664 0.9664 0.9664 149 0.8303 0.8457 0.8379 0.9940

Framework versions

  • Transformers 4.23.0.dev0
  • Pytorch 1.12.1+cu113
  • Datasets 2.2.2
  • Tokenizers 0.13.1
Downloads last month
8
Inference Examples
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.