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: 0.6791
  • Answer: {'precision': 0.6752411575562701, 'recall': 0.7787391841779975, 'f1': 0.7233065442020666, 'number': 809}
  • Header: {'precision': 0.25742574257425743, 'recall': 0.2184873949579832, 'f1': 0.23636363636363636, 'number': 119}
  • Question: {'precision': 0.7172995780590717, 'recall': 0.7981220657276995, 'f1': 0.7555555555555554, 'number': 1065}
  • Overall Precision: 0.6787
  • Overall Recall: 0.7556
  • Overall F1: 0.7151
  • Overall Accuracy: 0.7962

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: 3e-05
  • train_batch_size: 32
  • eval_batch_size: 16
  • seed: 42
  • optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
  • lr_scheduler_type: linear
  • num_epochs: 15
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Answer Header Question Overall Precision Overall Recall Overall F1 Overall Accuracy
1.8311 1.0 5 1.7018 {'precision': 0.015086206896551725, 'recall': 0.02595797280593325, 'f1': 0.01908223534756929, 'number': 809} {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119} {'precision': 0.09192692987625221, 'recall': 0.14647887323943662, 'f1': 0.11296162201303404, 'number': 1065} 0.0573 0.0888 0.0696 0.3364
1.6261 2.0 10 1.5278 {'precision': 0.018244013683010263, 'recall': 0.019777503090234856, 'f1': 0.018979833926453145, 'number': 809} {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119} {'precision': 0.24725943970767356, 'recall': 0.19061032863849764, 'f1': 0.21527041357370094, 'number': 1065} 0.1290 0.1099 0.1187 0.4110
1.4654 3.0 15 1.3491 {'precision': 0.07093023255813953, 'recall': 0.0754017305315204, 'f1': 0.07309766327142002, 'number': 809} {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119} {'precision': 0.3611111111111111, 'recall': 0.3539906103286385, 'f1': 0.35751541014698907, 'number': 1065} 0.2300 0.2198 0.2248 0.5293
1.2722 4.0 20 1.1745 {'precision': 0.2922222222222222, 'recall': 0.32509270704573545, 'f1': 0.30778232884727913, 'number': 809} {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119} {'precision': 0.4624, 'recall': 0.5427230046948357, 'f1': 0.49935205183585313, 'number': 1065} 0.3901 0.4220 0.4054 0.6268
1.0874 5.0 25 1.0226 {'precision': 0.4374331550802139, 'recall': 0.5055624227441285, 'f1': 0.4690366972477064, 'number': 809} {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119} {'precision': 0.5391849529780565, 'recall': 0.6460093896713615, 'f1': 0.5877829987184965, 'number': 1065} 0.4897 0.5504 0.5183 0.6874
0.9491 6.0 30 0.8969 {'precision': 0.5340022296544036, 'recall': 0.5920889987639061, 'f1': 0.5615474794841734, 'number': 809} {'precision': 0.07317073170731707, 'recall': 0.025210084033613446, 'f1': 0.0375, 'number': 119} {'precision': 0.6014376996805112, 'recall': 0.7070422535211267, 'f1': 0.6499784203711697, 'number': 1065} 0.5639 0.6197 0.5905 0.7330
0.8302 7.0 35 0.8232 {'precision': 0.5977482088024565, 'recall': 0.7218788627935723, 'f1': 0.6539753639417694, 'number': 809} {'precision': 0.1568627450980392, 'recall': 0.06722689075630252, 'f1': 0.09411764705882353, 'number': 119} {'precision': 0.6558669001751314, 'recall': 0.7032863849765258, 'f1': 0.678749433620299, 'number': 1065} 0.6180 0.6729 0.6442 0.7457
0.7414 8.0 40 0.7707 {'precision': 0.6148300720906282, 'recall': 0.7379480840543882, 'f1': 0.6707865168539326, 'number': 809} {'precision': 0.18333333333333332, 'recall': 0.09243697478991597, 'f1': 0.12290502793296088, 'number': 119} {'precision': 0.6633825944170771, 'recall': 0.7586854460093897, 'f1': 0.7078405606657906, 'number': 1065} 0.6296 0.7105 0.6676 0.7665
0.671 9.0 45 0.7335 {'precision': 0.6334012219959266, 'recall': 0.7688504326328801, 'f1': 0.6945840312674483, 'number': 809} {'precision': 0.2159090909090909, 'recall': 0.15966386554621848, 'f1': 0.18357487922705312, 'number': 119} {'precision': 0.6922413793103448, 'recall': 0.7539906103286385, 'f1': 0.7217977528089887, 'number': 1065} 0.6475 0.7245 0.6839 0.7742
0.6278 10.0 50 0.7206 {'precision': 0.649364406779661, 'recall': 0.757725587144623, 'f1': 0.6993725042783799, 'number': 809} {'precision': 0.21212121212121213, 'recall': 0.17647058823529413, 'f1': 0.1926605504587156, 'number': 119} {'precision': 0.6996587030716723, 'recall': 0.7699530516431925, 'f1': 0.7331247206079572, 'number': 1065} 0.6564 0.7296 0.6911 0.7847
0.5974 11.0 55 0.7095 {'precision': 0.653276955602537, 'recall': 0.7639060568603214, 'f1': 0.7042735042735042, 'number': 809} {'precision': 0.21505376344086022, 'recall': 0.16806722689075632, 'f1': 0.18867924528301888, 'number': 119} {'precision': 0.7091531223267751, 'recall': 0.7784037558685446, 'f1': 0.7421665174574755, 'number': 1065} 0.6644 0.7361 0.6984 0.7852
0.5594 12.0 60 0.6868 {'precision': 0.6523076923076923, 'recall': 0.7861557478368356, 'f1': 0.7130044843049326, 'number': 809} {'precision': 0.2619047619047619, 'recall': 0.18487394957983194, 'f1': 0.21674876847290642, 'number': 119} {'precision': 0.7068376068376069, 'recall': 0.7765258215962442, 'f1': 0.7400447427293065, 'number': 1065} 0.6662 0.7451 0.7035 0.7909
0.5374 13.0 65 0.6797 {'precision': 0.655958549222798, 'recall': 0.7824474660074165, 'f1': 0.7136414881623451, 'number': 809} {'precision': 0.25842696629213485, 'recall': 0.19327731092436976, 'f1': 0.22115384615384615, 'number': 119} {'precision': 0.7089678510998308, 'recall': 0.7868544600938967, 'f1': 0.7458834000890076, 'number': 1065} 0.6682 0.7496 0.7066 0.7926
0.5196 14.0 70 0.6794 {'precision': 0.673469387755102, 'recall': 0.7750309023485785, 'f1': 0.7206896551724138, 'number': 809} {'precision': 0.2631578947368421, 'recall': 0.21008403361344538, 'f1': 0.23364485981308414, 'number': 119} {'precision': 0.7148864592094197, 'recall': 0.7981220657276995, 'f1': 0.7542147293700088, 'number': 1065} 0.6781 0.7536 0.7139 0.7954
0.5065 15.0 75 0.6791 {'precision': 0.6752411575562701, 'recall': 0.7787391841779975, 'f1': 0.7233065442020666, 'number': 809} {'precision': 0.25742574257425743, 'recall': 0.2184873949579832, 'f1': 0.23636363636363636, 'number': 119} {'precision': 0.7172995780590717, 'recall': 0.7981220657276995, 'f1': 0.7555555555555554, 'number': 1065} 0.6787 0.7556 0.7151 0.7962

Framework versions

  • Transformers 4.47.0
  • Pytorch 2.5.1+cu121
  • Datasets 3.2.0
  • Tokenizers 0.21.0
Downloads last month
68
Safetensors
Model size
113M params
Tensor type
F32
·
Inference Providers NEW
This model is not currently available via any of the supported third-party Inference Providers, and the model is not deployed on the HF Inference API.

Model tree for thanhap03/layoutlm-funsd

Finetuned
(149)
this model