2024-01-07_one_stage_subgraphs_entropyreg_txt_vis_conc_2_5_9_11_gate
This model is a fine-tuned version of microsoft/layoutlmv3-base on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.9973
- Accuracy: 0.7725
- Exit 0 Accuracy: 0.0625
- Exit 1 Accuracy: 0.0625
- Exit 2 Accuracy: 0.0925
- Exit 3 Accuracy: 0.14
- Exit 4 Accuracy: 0.695
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: 2e-05
- train_batch_size: 2
- eval_batch_size: 1
- seed: 42
- gradient_accumulation_steps: 24
- total_train_batch_size: 48
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 60
Training results
Training Loss | Epoch | Step | Validation Loss | Accuracy | Exit 0 Accuracy | Exit 1 Accuracy | Exit 2 Accuracy | Exit 3 Accuracy | Exit 4 Accuracy |
---|---|---|---|---|---|---|---|---|---|
No log | 0.96 | 16 | 2.6595 | 0.1775 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 |
No log | 1.98 | 33 | 2.4912 | 0.27 | 0.065 | 0.0625 | 0.0625 | 0.0625 | 0.0625 |
No log | 3.0 | 50 | 2.2963 | 0.305 | 0.065 | 0.0625 | 0.0625 | 0.0625 | 0.0625 |
No log | 3.96 | 66 | 2.1349 | 0.38 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 |
No log | 4.98 | 83 | 1.9142 | 0.5275 | 0.065 | 0.0625 | 0.0625 | 0.0625 | 0.0625 |
No log | 6.0 | 100 | 1.6721 | 0.59 | 0.065 | 0.0625 | 0.0625 | 0.0625 | 0.0625 |
No log | 6.96 | 116 | 1.4696 | 0.6425 | 0.065 | 0.0625 | 0.0625 | 0.0625 | 0.0625 |
No log | 7.98 | 133 | 1.4285 | 0.665 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 |
No log | 9.0 | 150 | 1.2804 | 0.68 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0675 |
No log | 9.96 | 166 | 1.2030 | 0.7275 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0875 |
No log | 10.98 | 183 | 1.1252 | 0.73 | 0.065 | 0.0625 | 0.0625 | 0.0625 | 0.1175 |
No log | 12.0 | 200 | 1.0535 | 0.7375 | 0.065 | 0.0625 | 0.0625 | 0.0625 | 0.1275 |
No log | 12.96 | 216 | 1.0563 | 0.7375 | 0.065 | 0.0625 | 0.0625 | 0.0625 | 0.14 |
No log | 13.98 | 233 | 1.0299 | 0.7525 | 0.065 | 0.0625 | 0.0625 | 0.0625 | 0.155 |
No log | 15.0 | 250 | 0.9986 | 0.745 | 0.065 | 0.0625 | 0.0625 | 0.0625 | 0.17 |
No log | 15.96 | 266 | 1.0189 | 0.7525 | 0.065 | 0.0625 | 0.0625 | 0.0625 | 0.18 |
No log | 16.98 | 283 | 0.9661 | 0.7675 | 0.065 | 0.0625 | 0.0625 | 0.0625 | 0.19 |
No log | 18.0 | 300 | 0.9849 | 0.7475 | 0.065 | 0.0625 | 0.065 | 0.0625 | 0.195 |
No log | 18.96 | 316 | 0.9897 | 0.7325 | 0.0625 | 0.0625 | 0.065 | 0.0625 | 0.21 |
No log | 19.98 | 333 | 0.9983 | 0.7375 | 0.0625 | 0.0625 | 0.065 | 0.0625 | 0.2075 |
No log | 21.0 | 350 | 0.9799 | 0.7475 | 0.06 | 0.0625 | 0.0725 | 0.0625 | 0.21 |
No log | 21.96 | 366 | 0.9946 | 0.755 | 0.06 | 0.0625 | 0.065 | 0.065 | 0.2425 |
No log | 22.98 | 383 | 0.9856 | 0.75 | 0.0575 | 0.0625 | 0.07 | 0.0625 | 0.3125 |
No log | 24.0 | 400 | 0.9641 | 0.75 | 0.0625 | 0.0625 | 0.0825 | 0.0625 | 0.3225 |
No log | 24.96 | 416 | 0.9509 | 0.775 | 0.0575 | 0.0625 | 0.0825 | 0.0775 | 0.3625 |
No log | 25.98 | 433 | 0.9626 | 0.7775 | 0.0575 | 0.0625 | 0.0775 | 0.0875 | 0.4075 |
No log | 27.0 | 450 | 1.0076 | 0.7525 | 0.0575 | 0.0625 | 0.0825 | 0.0975 | 0.435 |
No log | 27.96 | 466 | 0.9588 | 0.765 | 0.06 | 0.0625 | 0.08 | 0.0975 | 0.4725 |
No log | 28.98 | 483 | 0.9423 | 0.765 | 0.06 | 0.0625 | 0.0775 | 0.1025 | 0.4725 |
0.6008 | 30.0 | 500 | 1.0173 | 0.7475 | 0.06 | 0.0625 | 0.08 | 0.1075 | 0.4675 |
0.6008 | 30.96 | 516 | 1.0115 | 0.7575 | 0.0625 | 0.0625 | 0.09 | 0.1175 | 0.505 |
0.6008 | 31.98 | 533 | 0.9999 | 0.765 | 0.0625 | 0.0625 | 0.1025 | 0.115 | 0.5575 |
0.6008 | 33.0 | 550 | 1.0163 | 0.77 | 0.06 | 0.0625 | 0.1025 | 0.135 | 0.6525 |
0.6008 | 33.96 | 566 | 0.9782 | 0.765 | 0.06 | 0.0625 | 0.1125 | 0.1225 | 0.62 |
0.6008 | 34.98 | 583 | 0.9811 | 0.7625 | 0.06 | 0.0625 | 0.1025 | 0.1425 | 0.6725 |
0.6008 | 36.0 | 600 | 0.9937 | 0.765 | 0.06 | 0.0625 | 0.1075 | 0.155 | 0.7 |
0.6008 | 36.96 | 616 | 1.0273 | 0.7575 | 0.06 | 0.0625 | 0.105 | 0.15 | 0.705 |
0.6008 | 37.98 | 633 | 1.0098 | 0.76 | 0.0625 | 0.0625 | 0.1075 | 0.1525 | 0.7 |
0.6008 | 39.0 | 650 | 0.9914 | 0.765 | 0.06 | 0.0625 | 0.1075 | 0.1325 | 0.6275 |
0.6008 | 39.96 | 666 | 1.0228 | 0.76 | 0.06 | 0.0625 | 0.1075 | 0.1175 | 0.605 |
0.6008 | 40.98 | 683 | 0.9701 | 0.7775 | 0.0575 | 0.0625 | 0.0925 | 0.1125 | 0.6175 |
0.6008 | 42.0 | 700 | 1.0100 | 0.7525 | 0.0575 | 0.0625 | 0.105 | 0.115 | 0.6225 |
0.6008 | 42.96 | 716 | 0.9989 | 0.77 | 0.06 | 0.0625 | 0.095 | 0.1325 | 0.6825 |
0.6008 | 43.98 | 733 | 0.9843 | 0.765 | 0.06 | 0.0625 | 0.1 | 0.1325 | 0.6825 |
0.6008 | 45.0 | 750 | 1.0016 | 0.7625 | 0.06 | 0.0625 | 0.0975 | 0.13 | 0.6525 |
0.6008 | 45.96 | 766 | 0.9942 | 0.765 | 0.06 | 0.0625 | 0.1 | 0.12 | 0.6775 |
0.6008 | 46.98 | 783 | 0.9977 | 0.7725 | 0.06 | 0.0625 | 0.0925 | 0.135 | 0.675 |
0.6008 | 48.0 | 800 | 1.0038 | 0.7725 | 0.0575 | 0.0625 | 0.09 | 0.135 | 0.6775 |
0.6008 | 48.96 | 816 | 1.0096 | 0.76 | 0.06 | 0.0625 | 0.0925 | 0.1225 | 0.6825 |
0.6008 | 49.98 | 833 | 0.9817 | 0.7725 | 0.06 | 0.0625 | 0.0925 | 0.135 | 0.7025 |
0.6008 | 51.0 | 850 | 0.9871 | 0.775 | 0.0625 | 0.0625 | 0.0925 | 0.145 | 0.705 |
0.6008 | 51.96 | 866 | 1.0015 | 0.765 | 0.0625 | 0.0625 | 0.0925 | 0.1425 | 0.685 |
0.6008 | 52.98 | 883 | 0.9924 | 0.77 | 0.0625 | 0.0625 | 0.0925 | 0.14 | 0.6975 |
0.6008 | 54.0 | 900 | 1.0002 | 0.765 | 0.0625 | 0.0625 | 0.0925 | 0.1375 | 0.6975 |
0.6008 | 54.96 | 916 | 0.9995 | 0.7775 | 0.0625 | 0.0625 | 0.0925 | 0.1375 | 0.7 |
0.6008 | 55.98 | 933 | 0.9971 | 0.77 | 0.0625 | 0.0625 | 0.0925 | 0.14 | 0.6975 |
0.6008 | 57.0 | 950 | 0.9973 | 0.7725 | 0.0625 | 0.0625 | 0.0925 | 0.14 | 0.6925 |
0.6008 | 57.6 | 960 | 0.9973 | 0.7725 | 0.0625 | 0.0625 | 0.0925 | 0.14 | 0.695 |
Framework versions
- Transformers 4.31.0
- Pytorch 2.0.1+cu117
- Datasets 2.13.1
- Tokenizers 0.13.3
- Downloads last month
- 1