<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->
EElayoutlmv3_jordyvl_rvl_cdip_100_examples_per_class_2023-09-26_subgraphs_gates_exitloss
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: 1.6039
- Accuracy: 0.7225
- Exit 0 Accuracy: 0.1275
- Exit 1 Accuracy: 0.125
- Exit 2 Accuracy: 0.5375
- Exit 3 Accuracy: 0.71
- Exit 4 Accuracy: 0.7175
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: 4
- eval_batch_size: 1
- seed: 42
- gradient_accumulation_steps: 12
- 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 | 2.6598 | 0.1675 | 0.065 | 0.0975 | 0.0875 | 0.0625 | 0.0625 |
No log | 1.96 | 32 | 2.5926 | 0.22 | 0.0675 | 0.105 | 0.1 | 0.125 | 0.14 |
No log | 2.96 | 48 | 2.3432 | 0.345 | 0.08 | 0.115 | 0.17 | 0.26 | 0.2925 |
No log | 3.96 | 64 | 2.1196 | 0.4125 | 0.08 | 0.1125 | 0.2475 | 0.405 | 0.3925 |
No log | 4.96 | 80 | 1.8286 | 0.47 | 0.09 | 0.11 | 0.285 | 0.485 | 0.5275 |
No log | 5.96 | 96 | 1.6371 | 0.535 | 0.0975 | 0.11 | 0.2975 | 0.5275 | 0.53 |
No log | 6.96 | 112 | 1.4443 | 0.6025 | 0.1 | 0.1125 | 0.3325 | 0.5725 | 0.585 |
No log | 7.96 | 128 | 1.3876 | 0.595 | 0.105 | 0.11 | 0.3675 | 0.585 | 0.615 |
No log | 8.96 | 144 | 1.3300 | 0.6175 | 0.1125 | 0.11 | 0.375 | 0.5925 | 0.6225 |
No log | 9.96 | 160 | 1.2474 | 0.6425 | 0.11 | 0.11 | 0.3725 | 0.6025 | 0.62 |
No log | 10.96 | 176 | 1.2704 | 0.6175 | 0.1025 | 0.11 | 0.3725 | 0.6125 | 0.645 |
No log | 11.96 | 192 | 1.2148 | 0.65 | 0.11 | 0.11 | 0.38 | 0.6075 | 0.64 |
No log | 12.96 | 208 | 1.1749 | 0.6825 | 0.105 | 0.1125 | 0.39 | 0.625 | 0.665 |
No log | 13.96 | 224 | 1.1786 | 0.6675 | 0.1025 | 0.1125 | 0.385 | 0.655 | 0.6775 |
No log | 14.96 | 240 | 1.1414 | 0.6875 | 0.1175 | 0.1125 | 0.3975 | 0.645 | 0.6775 |
No log | 15.96 | 256 | 1.1387 | 0.695 | 0.1125 | 0.1125 | 0.3925 | 0.645 | 0.69 |
No log | 16.96 | 272 | 1.2557 | 0.67 | 0.1275 | 0.1125 | 0.405 | 0.635 | 0.6625 |
No log | 17.96 | 288 | 1.2320 | 0.6825 | 0.1125 | 0.1125 | 0.41 | 0.6525 | 0.6725 |
No log | 18.96 | 304 | 1.2218 | 0.6925 | 0.1275 | 0.1125 | 0.43 | 0.66 | 0.6975 |
No log | 19.96 | 320 | 1.2435 | 0.6825 | 0.1275 | 0.1125 | 0.45 | 0.66 | 0.69 |
No log | 20.96 | 336 | 1.2608 | 0.6975 | 0.125 | 0.11 | 0.4475 | 0.67 | 0.69 |
No log | 21.96 | 352 | 1.2359 | 0.7025 | 0.135 | 0.11 | 0.4675 | 0.675 | 0.7075 |
No log | 22.96 | 368 | 1.2474 | 0.7125 | 0.135 | 0.1125 | 0.4575 | 0.68 | 0.705 |
No log | 23.96 | 384 | 1.2588 | 0.7125 | 0.125 | 0.1125 | 0.4575 | 0.6725 | 0.715 |
No log | 24.96 | 400 | 1.3241 | 0.7025 | 0.115 | 0.1125 | 0.45 | 0.665 | 0.695 |
No log | 25.96 | 416 | 1.3844 | 0.695 | 0.1275 | 0.1125 | 0.465 | 0.6775 | 0.6925 |
No log | 26.96 | 432 | 1.3304 | 0.72 | 0.125 | 0.1125 | 0.475 | 0.6875 | 0.7075 |
No log | 27.96 | 448 | 1.3437 | 0.7025 | 0.1425 | 0.1125 | 0.4825 | 0.6925 | 0.71 |
No log | 28.96 | 464 | 1.3657 | 0.6975 | 0.135 | 0.1125 | 0.47 | 0.6925 | 0.7125 |
No log | 29.96 | 480 | 1.4499 | 0.6825 | 0.155 | 0.1125 | 0.49 | 0.6875 | 0.7 |
No log | 30.96 | 496 | 1.4276 | 0.6975 | 0.1325 | 0.1125 | 0.49 | 0.6925 | 0.7025 |
13.2498 | 31.96 | 512 | 1.4664 | 0.71 | 0.1325 | 0.1125 | 0.475 | 0.69 | 0.71 |
13.2498 | 32.96 | 528 | 1.4442 | 0.7125 | 0.12 | 0.1125 | 0.495 | 0.6975 | 0.705 |
13.2498 | 33.96 | 544 | 1.4387 | 0.725 | 0.1125 | 0.115 | 0.4975 | 0.7075 | 0.725 |
13.2498 | 34.96 | 560 | 1.4703 | 0.7225 | 0.1125 | 0.115 | 0.4875 | 0.7 | 0.72 |
13.2498 | 35.96 | 576 | 1.4634 | 0.71 | 0.13 | 0.115 | 0.5 | 0.7025 | 0.72 |
13.2498 | 36.96 | 592 | 1.4800 | 0.715 | 0.14 | 0.1175 | 0.5 | 0.7 | 0.7125 |
13.2498 | 37.96 | 608 | 1.4798 | 0.7275 | 0.115 | 0.12 | 0.505 | 0.7125 | 0.73 |
13.2498 | 38.96 | 624 | 1.4689 | 0.72 | 0.1525 | 0.1225 | 0.5075 | 0.705 | 0.71 |
13.2498 | 39.96 | 640 | 1.5499 | 0.71 | 0.1125 | 0.1225 | 0.51 | 0.7075 | 0.7075 |
13.2498 | 40.96 | 656 | 1.4984 | 0.72 | 0.1275 | 0.1225 | 0.51 | 0.715 | 0.7175 |
13.2498 | 41.96 | 672 | 1.5532 | 0.705 | 0.1125 | 0.1225 | 0.51 | 0.71 | 0.71 |
13.2498 | 42.96 | 688 | 1.5209 | 0.7125 | 0.145 | 0.1225 | 0.53 | 0.7225 | 0.7225 |
13.2498 | 43.96 | 704 | 1.5401 | 0.715 | 0.1325 | 0.1225 | 0.5225 | 0.71 | 0.71 |
13.2498 | 44.96 | 720 | 1.5469 | 0.7275 | 0.135 | 0.1225 | 0.5075 | 0.71 | 0.7125 |
13.2498 | 45.96 | 736 | 1.5717 | 0.7225 | 0.13 | 0.1225 | 0.5325 | 0.7175 | 0.715 |
13.2498 | 46.96 | 752 | 1.5716 | 0.7175 | 0.1475 | 0.1225 | 0.5325 | 0.7125 | 0.72 |
13.2498 | 47.96 | 768 | 1.5648 | 0.7225 | 0.1225 | 0.1225 | 0.5275 | 0.7125 | 0.7175 |
13.2498 | 48.96 | 784 | 1.5605 | 0.725 | 0.1275 | 0.125 | 0.53 | 0.72 | 0.725 |
13.2498 | 49.96 | 800 | 1.5793 | 0.72 | 0.12 | 0.125 | 0.5325 | 0.7225 | 0.7175 |
13.2498 | 50.96 | 816 | 1.5791 | 0.72 | 0.1275 | 0.125 | 0.5425 | 0.72 | 0.7175 |
13.2498 | 51.96 | 832 | 1.5875 | 0.715 | 0.1325 | 0.125 | 0.535 | 0.7125 | 0.72 |
13.2498 | 52.96 | 848 | 1.6052 | 0.7175 | 0.1225 | 0.125 | 0.54 | 0.7075 | 0.715 |
13.2498 | 53.96 | 864 | 1.5907 | 0.7225 | 0.1275 | 0.125 | 0.54 | 0.7125 | 0.7225 |
13.2498 | 54.96 | 880 | 1.5928 | 0.73 | 0.13 | 0.125 | 0.5375 | 0.7075 | 0.7175 |
13.2498 | 55.96 | 896 | 1.5986 | 0.72 | 0.13 | 0.125 | 0.5325 | 0.705 | 0.7175 |
13.2498 | 56.96 | 912 | 1.5946 | 0.725 | 0.125 | 0.125 | 0.535 | 0.7075 | 0.7225 |
13.2498 | 57.96 | 928 | 1.5985 | 0.725 | 0.125 | 0.125 | 0.54 | 0.71 | 0.715 |
13.2498 | 58.96 | 944 | 1.6036 | 0.7225 | 0.1275 | 0.125 | 0.5375 | 0.71 | 0.7175 |
13.2498 | 59.96 | 960 | 1.6039 | 0.7225 | 0.1275 | 0.125 | 0.5375 | 0.71 | 0.7175 |
Framework versions
- Transformers 4.26.1
- Pytorch 1.13.1.post200
- Datasets 2.9.0
- Tokenizers 0.13.2