2024-01-08_one_stage_subgraphs_weighted_entropyreg_txt_vision_enc_all_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: 0.9591
- Accuracy: 0.77
- Exit 0 Accuracy: 0.055
- Exit 1 Accuracy: 0.0675
- Exit 2 Accuracy: 0.0625
- Exit 3 Accuracy: 0.0625
- Exit 4 Accuracy: 0.0625
- Exit 5 Accuracy: 0.0625
- Exit 6 Accuracy: 0.0625
- Exit 7 Accuracy: 0.0625
- Exit 8 Accuracy: 0.0625
- Exit 9 Accuracy: 0.0625
- Exit 10 Accuracy: 0.0625
- Exit 11 Accuracy: 0.0625
- Exit 12 Accuracy: 0.5325
- Exit 13 Accuracy: 0.77
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 | Exit 5 Accuracy | Exit 6 Accuracy | Exit 7 Accuracy | Exit 8 Accuracy | Exit 9 Accuracy | Exit 10 Accuracy | Exit 11 Accuracy | Exit 12 Accuracy | Exit 13 Accuracy |
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
No log | 0.96 | 16 | 2.7132 | 0.1375 | 0.0425 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.1375 |
No log | 1.98 | 33 | 2.5541 | 0.225 | 0.04 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.225 |
No log | 3.0 | 50 | 2.3875 | 0.2925 | 0.06 | 0.0675 | 0.06 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.2925 |
No log | 3.96 | 66 | 2.2225 | 0.355 | 0.055 | 0.0675 | 0.06 | 0.0625 | 0.0625 | 0.0625 | 0.065 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.355 |
No log | 4.98 | 83 | 2.0628 | 0.4225 | 0.0625 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.4225 |
No log | 6.0 | 100 | 1.8932 | 0.49 | 0.0625 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.49 |
No log | 6.96 | 116 | 1.6240 | 0.5725 | 0.0625 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.5725 |
No log | 7.98 | 133 | 1.4324 | 0.6575 | 0.0625 | 0.0675 | 0.065 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.6575 |
No log | 9.0 | 150 | 1.2942 | 0.7 | 0.0625 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.7 |
No log | 9.96 | 166 | 1.1669 | 0.75 | 0.0625 | 0.0675 | 0.065 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.75 |
No log | 10.98 | 183 | 1.1042 | 0.7475 | 0.0625 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.7475 |
No log | 12.0 | 200 | 1.0280 | 0.7425 | 0.06 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.7425 |
No log | 12.96 | 216 | 1.0155 | 0.7325 | 0.06 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.7325 |
No log | 13.98 | 233 | 1.0058 | 0.745 | 0.06 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.065 | 0.745 |
No log | 15.0 | 250 | 0.9816 | 0.74 | 0.06 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.74 |
No log | 15.96 | 266 | 0.9253 | 0.765 | 0.0625 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.065 | 0.765 |
No log | 16.98 | 283 | 1.0265 | 0.715 | 0.06 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0775 | 0.715 |
No log | 18.0 | 300 | 0.9361 | 0.7575 | 0.0625 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.075 | 0.7575 |
No log | 18.96 | 316 | 0.9585 | 0.7475 | 0.0575 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0825 | 0.7475 |
No log | 19.98 | 333 | 0.9825 | 0.755 | 0.0575 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.1125 | 0.755 |
No log | 21.0 | 350 | 0.9122 | 0.77 | 0.0675 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.13 | 0.77 |
No log | 21.96 | 366 | 0.9313 | 0.7675 | 0.0575 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.1375 | 0.7675 |
No log | 22.98 | 383 | 0.9270 | 0.7725 | 0.0475 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.1625 | 0.7725 |
No log | 24.0 | 400 | 0.9181 | 0.7675 | 0.065 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.1675 | 0.7675 |
No log | 24.96 | 416 | 0.9614 | 0.7625 | 0.065 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.18 | 0.7625 |
No log | 25.98 | 433 | 0.9338 | 0.7775 | 0.06 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.205 | 0.7775 |
No log | 27.0 | 450 | 0.9171 | 0.7775 | 0.0625 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.26 | 0.7775 |
No log | 27.96 | 466 | 0.9374 | 0.77 | 0.0625 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.285 | 0.77 |
No log | 28.98 | 483 | 0.9243 | 0.77 | 0.065 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.32 | 0.77 |
0.4492 | 30.0 | 500 | 0.9471 | 0.77 | 0.065 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.3175 | 0.77 |
0.4492 | 30.96 | 516 | 0.9419 | 0.7675 | 0.065 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.35 | 0.7675 |
0.4492 | 31.98 | 533 | 0.9573 | 0.7625 | 0.065 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.3675 | 0.7625 |
0.4492 | 33.0 | 550 | 0.9631 | 0.7625 | 0.065 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.37 | 0.7625 |
0.4492 | 33.96 | 566 | 0.9474 | 0.7675 | 0.065 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.3875 | 0.7675 |
0.4492 | 34.98 | 583 | 0.9661 | 0.765 | 0.065 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.39 | 0.765 |
0.4492 | 36.0 | 600 | 0.9507 | 0.77 | 0.065 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.395 | 0.77 |
0.4492 | 36.96 | 616 | 0.9496 | 0.77 | 0.065 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.4075 | 0.77 |
0.4492 | 37.98 | 633 | 0.9516 | 0.765 | 0.0625 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.4175 | 0.765 |
0.4492 | 39.0 | 650 | 0.9560 | 0.7675 | 0.0575 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.42 | 0.7675 |
0.4492 | 39.96 | 666 | 0.9453 | 0.775 | 0.0475 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.4225 | 0.775 |
0.4492 | 40.98 | 683 | 0.9468 | 0.775 | 0.05 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.4275 | 0.775 |
0.4492 | 42.0 | 700 | 0.9584 | 0.77 | 0.05 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.4375 | 0.77 |
0.4492 | 42.96 | 716 | 0.9623 | 0.7625 | 0.0525 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.46 | 0.7625 |
0.4492 | 43.98 | 733 | 0.9505 | 0.7625 | 0.0525 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.4725 | 0.7625 |
0.4492 | 45.0 | 750 | 0.9514 | 0.7725 | 0.055 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.4825 | 0.7725 |
0.4492 | 45.96 | 766 | 0.9509 | 0.7725 | 0.055 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.4925 | 0.7725 |
0.4492 | 46.98 | 783 | 0.9542 | 0.775 | 0.055 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.4925 | 0.775 |
0.4492 | 48.0 | 800 | 0.9545 | 0.7675 | 0.055 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.4975 | 0.7675 |
0.4492 | 48.96 | 816 | 0.9582 | 0.775 | 0.055 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.4975 | 0.775 |
0.4492 | 49.98 | 833 | 0.9563 | 0.765 | 0.055 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.5 | 0.765 |
0.4492 | 51.0 | 850 | 0.9580 | 0.7675 | 0.055 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.5125 | 0.7675 |
0.4492 | 51.96 | 866 | 0.9609 | 0.7625 | 0.055 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.515 | 0.7625 |
0.4492 | 52.98 | 883 | 0.9610 | 0.7625 | 0.055 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.51 | 0.7625 |
0.4492 | 54.0 | 900 | 0.9591 | 0.7625 | 0.055 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.52 | 0.7625 |
0.4492 | 54.96 | 916 | 0.9565 | 0.77 | 0.055 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.5275 | 0.77 |
0.4492 | 55.98 | 933 | 0.9591 | 0.77 | 0.055 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.5325 | 0.77 |
0.4492 | 57.0 | 950 | 0.9591 | 0.77 | 0.055 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.5325 | 0.77 |
0.4492 | 57.6 | 960 | 0.9591 | 0.77 | 0.055 | 0.0675 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.0625 | 0.5325 | 0.77 |
Framework versions
- Transformers 4.31.0
- Pytorch 2.0.1+cu117
- Datasets 2.13.1
- Tokenizers 0.13.3
- Downloads last month
- 0