metadata
license: mit
base_model: microsoft/layoutlm-base-uncased
tags:
- generated_from_trainer
datasets:
- funsd
model-index:
- name: layoutlm-funsd3
results: []
layoutlm-funsd3
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: 1.1833
- Answer: {'precision': 0.2526041666666667, 'recall': 0.23980222496909764, 'f1': 0.24603677869372226, 'number': 809}
- Header: {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119}
- Question: {'precision': 0.4935064935064935, 'recall': 0.5352112676056338, 'f1': 0.5135135135135136, 'number': 1065}
- Overall Precision: 0.3973
- Overall Recall: 0.3833
- Overall F1: 0.3902
- Overall Accuracy: 0.6048
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: 16
- eval_batch_size: 8
- 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: 12
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Answer | Header | Question | Overall Precision | Overall Recall | Overall F1 | Overall Accuracy |
|---|---|---|---|---|---|---|---|---|---|---|
| 1.9597 | 1.0 | 10 | 1.9692 | {'precision': 0.024831867563372995, 'recall': 0.059332509270704575, 'f1': 0.0350109409190372, 'number': 809} | {'precision': 0.0054894784995425435, 'recall': 0.05042016806722689, 'f1': 0.009900990099009901, 'number': 119} | {'precision': 0.05862516212710765, 'recall': 0.21220657276995306, 'f1': 0.091869918699187, 'number': 1065} | 0.0407 | 0.1405 | 0.0631 | 0.1655 |
| 1.9429 | 2.0 | 20 | 1.9517 | {'precision': 0.02355889724310777, 'recall': 0.0580964153275649, 'f1': 0.033523537803138374, 'number': 809} | {'precision': 0.006984866123399301, 'recall': 0.05042016806722689, 'f1': 0.012269938650306747, 'number': 119} | {'precision': 0.06357435197817189, 'recall': 0.2187793427230047, 'f1': 0.09852008456659618, 'number': 1065} | 0.0439 | 0.1435 | 0.0672 | 0.1837 |
| 1.9283 | 3.0 | 30 | 1.9222 | {'precision': 0.02506265664160401, 'recall': 0.06180469715698393, 'f1': 0.035663338088445073, 'number': 809} | {'precision': 0.005802707930367505, 'recall': 0.025210084033613446, 'f1': 0.009433962264150943, 'number': 119} | {'precision': 0.0683998761993191, 'recall': 0.20751173708920187, 'f1': 0.10288640595903166, 'number': 1065} | 0.0477 | 0.1375 | 0.0708 | 0.2076 |
| 1.8979 | 4.0 | 40 | 1.8822 | {'precision': 0.026082130965593784, 'recall': 0.0580964153275649, 'f1': 0.03600153198008425, 'number': 809} | {'precision': 0.01327433628318584, 'recall': 0.025210084033613446, 'f1': 0.017391304347826087, 'number': 119} | {'precision': 0.07303807303807304, 'recall': 0.17652582159624414, 'f1': 0.10332508931025007, 'number': 1065} | 0.0517 | 0.1194 | 0.0722 | 0.2412 |
| 1.8452 | 5.0 | 50 | 1.8330 | {'precision': 0.02404809619238477, 'recall': 0.04449938195302843, 'f1': 0.03122289679098005, 'number': 809} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119} | {'precision': 0.08863636363636364, 'recall': 0.14647887323943662, 'f1': 0.11044247787610618, 'number': 1065} | 0.0577 | 0.0963 | 0.0722 | 0.2719 |
| 1.7986 | 6.0 | 60 | 1.7759 | {'precision': 0.017241379310344827, 'recall': 0.022249690976514216, 'f1': 0.019427954668105776, 'number': 809} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119} | {'precision': 0.1244196843082637, 'recall': 0.12582159624413145, 'f1': 0.12511671335200747, 'number': 1065} | 0.0713 | 0.0763 | 0.0737 | 0.3012 |
| 1.7397 | 7.0 | 70 | 1.7107 | {'precision': 0.02045728038507822, 'recall': 0.021013597033374538, 'f1': 0.02073170731707317, 'number': 809} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119} | {'precision': 0.18026315789473685, 'recall': 0.12863849765258217, 'f1': 0.15013698630136987, 'number': 1065} | 0.0967 | 0.0773 | 0.0859 | 0.3270 |
| 1.6707 | 8.0 | 80 | 1.6298 | {'precision': 0.03066271018793274, 'recall': 0.038318912237330034, 'f1': 0.03406593406593406, 'number': 809} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119} | {'precision': 0.20087815587266739, 'recall': 0.17183098591549295, 'f1': 0.18522267206477733, 'number': 1065} | 0.1113 | 0.1074 | 0.1093 | 0.3654 |
| 1.5891 | 9.0 | 90 | 1.5416 | {'precision': 0.047619047619047616, 'recall': 0.06674907292954264, 'f1': 0.055584148224395266, 'number': 809} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119} | {'precision': 0.2087421944692239, 'recall': 0.21971830985915494, 'f1': 0.2140896614821592, 'number': 1065} | 0.1277 | 0.1445 | 0.1356 | 0.4183 |
| 1.516 | 10.0 | 100 | 1.4443 | {'precision': 0.06370070778564206, 'recall': 0.07787391841779975, 'f1': 0.07007786429365963, 'number': 809} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119} | {'precision': 0.2838983050847458, 'recall': 0.3145539906103286, 'f1': 0.2984409799554566, 'number': 1065} | 0.1835 | 0.1997 | 0.1913 | 0.4720 |
| 1.3887 | 11.0 | 110 | 1.3259 | {'precision': 0.11662531017369727, 'recall': 0.1161928306551298, 'f1': 0.11640866873065016, 'number': 809} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119} | {'precision': 0.39842381786339753, 'recall': 0.4272300469483568, 'f1': 0.412324422292705, 'number': 1065} | 0.2818 | 0.2755 | 0.2786 | 0.5434 |
| 1.261 | 12.0 | 120 | 1.1833 | {'precision': 0.2526041666666667, 'recall': 0.23980222496909764, 'f1': 0.24603677869372226, 'number': 809} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119} | {'precision': 0.4935064935064935, 'recall': 0.5352112676056338, 'f1': 0.5135135135135136, 'number': 1065} | 0.3973 | 0.3833 | 0.3902 | 0.6048 |
Framework versions
- Transformers 4.41.2
- Pytorch 2.3.1+cu121
- Datasets 2.19.2
- Tokenizers 0.19.1