layoutlm-funsd / README.md
Haios141's picture
End of training
ec6c57b
metadata
base_model: microsoft/layoutlm-base-uncased
tags:
  - generated_from_trainer
datasets:
  - funsd
model-index:
  - name: layoutlm-funsd
    results: []

layoutlm-funsd

This model is a fine-tuned version of microsoft/layoutlm-base-uncased on the funsd dataset. It achieves the following results on the evaluation set:

  • Loss: 1.3057
  • Answer: {'precision': 0.09480519480519481, 'recall': 0.09023485784919653, 'f1': 0.09246358454718177, 'number': 809}
  • Header: {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119}
  • Question: {'precision': 0.4032534246575342, 'recall': 0.4422535211267606, 'f1': 0.4218540080609046, 'number': 1065}
  • Overall Precision: 0.2807
  • Overall Recall: 0.2730
  • Overall F1: 0.2768
  • Overall Accuracy: 0.5691

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-06
  • train_batch_size: 16
  • eval_batch_size: 8
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 15

Training results

Training Loss Epoch Step Validation Loss Answer Header Question Overall Precision Overall Recall Overall F1 Overall Accuracy
1.9048 1.0 10 1.8492 {'precision': 0.02683982683982684, 'recall': 0.07663782447466007, 'f1': 0.039756332157742866, 'number': 809} {'precision': 0.003424657534246575, 'recall': 0.008403361344537815, 'f1': 0.004866180048661801, 'number': 119} {'precision': 0.08558262014483213, 'recall': 0.12206572769953052, 'f1': 0.10061919504643962, 'number': 1065} 0.0468 0.0968 0.0631 0.2625
1.8261 2.0 20 1.7805 {'precision': 0.02488425925925926, 'recall': 0.05315203955500618, 'f1': 0.03389830508474576, 'number': 809} {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119} {'precision': 0.11639344262295082, 'recall': 0.13333333333333333, 'f1': 0.12428884026258205, 'number': 1065} 0.0620 0.0928 0.0744 0.3314
1.7557 3.0 30 1.7197 {'precision': 0.018808777429467086, 'recall': 0.029666254635352288, 'f1': 0.02302158273381295, 'number': 809} {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119} {'precision': 0.15336134453781514, 'recall': 0.13708920187793427, 'f1': 0.14476945959345563, 'number': 1065} 0.0763 0.0853 0.0805 0.3579
1.7002 4.0 40 1.6648 {'precision': 0.019029495718363463, 'recall': 0.024721878862793572, 'f1': 0.02150537634408602, 'number': 809} {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119} {'precision': 0.19602977667493796, 'recall': 0.14835680751173708, 'f1': 0.16889363976483165, 'number': 1065} 0.0959 0.0893 0.0925 0.3775
1.645 5.0 50 1.6121 {'precision': 0.019801980198019802, 'recall': 0.024721878862793572, 'f1': 0.021990104452996154, 'number': 809} {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119} {'precision': 0.22172452407614782, 'recall': 0.18591549295774648, 'f1': 0.20224719101123598, 'number': 1065} 0.1146 0.1094 0.1119 0.4091
1.5951 6.0 60 1.5596 {'precision': 0.029411764705882353, 'recall': 0.037082818294190356, 'f1': 0.032804811372334604, 'number': 809} {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119} {'precision': 0.23694779116465864, 'recall': 0.2215962441314554, 'f1': 0.2290150412421155, 'number': 1065} 0.1319 0.1335 0.1327 0.4421
1.5418 7.0 70 1.5109 {'precision': 0.040755467196819085, 'recall': 0.05067985166872682, 'f1': 0.04517906336088154, 'number': 809} {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119} {'precision': 0.27926267281105993, 'recall': 0.28450704225352114, 'f1': 0.2818604651162791, 'number': 1065} 0.1645 0.1726 0.1685 0.4719
1.4954 8.0 80 1.4653 {'precision': 0.050359712230215826, 'recall': 0.06056860321384425, 'f1': 0.05499438832772166, 'number': 809} {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119} {'precision': 0.3016421780466724, 'recall': 0.3276995305164319, 'f1': 0.31413141314131415, 'number': 1065} 0.1869 0.1997 0.1931 0.4973
1.4558 9.0 90 1.4245 {'precision': 0.054140127388535034, 'recall': 0.0630407911001236, 'f1': 0.05825242718446602, 'number': 809} {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119} {'precision': 0.3177966101694915, 'recall': 0.352112676056338, 'f1': 0.3340757238307349, 'number': 1065} 0.2008 0.2137 0.2070 0.5168
1.4126 10.0 100 1.3893 {'precision': 0.07432432432432433, 'recall': 0.0815822002472188, 'f1': 0.07778432527990571, 'number': 809} {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119} {'precision': 0.33669185558354325, 'recall': 0.37652582159624415, 'f1': 0.3554964539007092, 'number': 1065} 0.2246 0.2343 0.2294 0.5339
1.3759 11.0 110 1.3592 {'precision': 0.08333333333333333, 'recall': 0.0865265760197775, 'f1': 0.08489993935718616, 'number': 809} {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119} {'precision': 0.3618807724601176, 'recall': 0.40469483568075115, 'f1': 0.38209219858156024, 'number': 1065} 0.2467 0.2514 0.2490 0.5470
1.3663 12.0 120 1.3358 {'precision': 0.08531994981179424, 'recall': 0.08405438813349815, 'f1': 0.08468244084682441, 'number': 809} {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119} {'precision': 0.37638062871707734, 'recall': 0.415962441314554, 'f1': 0.39518287243532557, 'number': 1065} 0.2589 0.2564 0.2576 0.5545
1.3323 13.0 130 1.3192 {'precision': 0.0916030534351145, 'recall': 0.08899876390605686, 'f1': 0.090282131661442, 'number': 809} {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119} {'precision': 0.38649789029535864, 'recall': 0.4300469483568075, 'f1': 0.40711111111111115, 'number': 1065} 0.2689 0.2659 0.2674 0.5635
1.3268 14.0 140 1.3094 {'precision': 0.09585492227979274, 'recall': 0.09147095179233622, 'f1': 0.09361163820366855, 'number': 809} {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119} {'precision': 0.3974358974358974, 'recall': 0.43661971830985913, 'f1': 0.4161073825503355, 'number': 1065} 0.2775 0.2704 0.2740 0.5671
1.2988 15.0 150 1.3057 {'precision': 0.09480519480519481, 'recall': 0.09023485784919653, 'f1': 0.09246358454718177, 'number': 809} {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119} {'precision': 0.4032534246575342, 'recall': 0.4422535211267606, 'f1': 0.4218540080609046, 'number': 1065} 0.2807 0.2730 0.2768 0.5691

Framework versions

  • Transformers 4.31.0
  • Pytorch 2.0.1+cu117
  • Datasets 2.14.4
  • Tokenizers 0.13.3