<!-- 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_2023-07-06_g085
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: 1.3237
- Accuracy: 0.6425
- Exit 0 Accuracy: 0.1125
- Exit 1 Accuracy: 0.1625
- Exit 2 Accuracy: 0.385
- Exit 3 Accuracy: 0.555
- Exit 4 Accuracy: 0.6075
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: 12
- eval_batch_size: 4
- seed: 42
- gradient_accumulation_steps: 24
- total_train_batch_size: 288
- 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 | 0.72 | 2 | 2.7611 | 0.1075 | 0.0875 | 0.0675 | 0.105 | 0.0625 | 0.0625 |
No log | 1.72 | 4 | 2.7343 | 0.1125 | 0.07 | 0.065 | 0.1275 | 0.0625 | 0.0625 |
No log | 2.72 | 6 | 2.6999 | 0.13 | 0.0775 | 0.06 | 0.1375 | 0.0625 | 0.0625 |
No log | 3.72 | 8 | 2.6641 | 0.165 | 0.08 | 0.06 | 0.1225 | 0.0625 | 0.0625 |
No log | 4.72 | 10 | 2.6228 | 0.1875 | 0.0875 | 0.0575 | 0.115 | 0.0625 | 0.0625 |
No log | 5.72 | 12 | 2.5767 | 0.205 | 0.085 | 0.08 | 0.125 | 0.08 | 0.0625 |
No log | 6.72 | 14 | 2.5435 | 0.2275 | 0.0925 | 0.0825 | 0.155 | 0.1075 | 0.0625 |
No log | 7.72 | 16 | 2.5258 | 0.24 | 0.1 | 0.095 | 0.1825 | 0.115 | 0.065 |
No log | 8.72 | 18 | 2.4842 | 0.28 | 0.1 | 0.0975 | 0.2 | 0.1225 | 0.0775 |
No log | 9.72 | 20 | 2.4249 | 0.285 | 0.1 | 0.1325 | 0.1975 | 0.16 | 0.075 |
No log | 10.72 | 22 | 2.3847 | 0.3075 | 0.115 | 0.13 | 0.21 | 0.175 | 0.065 |
No log | 11.72 | 24 | 2.3568 | 0.3075 | 0.115 | 0.1175 | 0.215 | 0.24 | 0.09 |
No log | 12.72 | 26 | 2.3226 | 0.31 | 0.1125 | 0.1275 | 0.21 | 0.2375 | 0.1075 |
No log | 13.72 | 28 | 2.2816 | 0.3325 | 0.115 | 0.125 | 0.2225 | 0.23 | 0.1075 |
No log | 14.72 | 30 | 2.2523 | 0.33 | 0.1125 | 0.1325 | 0.2325 | 0.2575 | 0.11 |
No log | 15.72 | 32 | 2.2313 | 0.33 | 0.1175 | 0.135 | 0.235 | 0.2825 | 0.1125 |
No log | 16.72 | 34 | 2.1980 | 0.3425 | 0.1125 | 0.1375 | 0.26 | 0.31 | 0.11 |
No log | 17.72 | 36 | 2.1571 | 0.345 | 0.1075 | 0.1375 | 0.2625 | 0.32 | 0.1125 |
No log | 18.72 | 38 | 2.1287 | 0.35 | 0.11 | 0.14 | 0.2925 | 0.3525 | 0.1125 |
No log | 19.72 | 40 | 2.1027 | 0.36 | 0.11 | 0.14 | 0.315 | 0.3525 | 0.1075 |
No log | 20.72 | 42 | 2.0783 | 0.3925 | 0.11 | 0.1425 | 0.32 | 0.33 | 0.1325 |
No log | 21.72 | 44 | 2.0591 | 0.395 | 0.1125 | 0.14 | 0.3075 | 0.2975 | 0.1925 |
No log | 22.72 | 46 | 2.0173 | 0.4175 | 0.1125 | 0.14 | 0.31 | 0.2975 | 0.2075 |
No log | 23.72 | 48 | 1.9916 | 0.4275 | 0.1125 | 0.1425 | 0.32 | 0.3225 | 0.2175 |
No log | 24.72 | 50 | 1.9750 | 0.4375 | 0.1125 | 0.1425 | 0.3225 | 0.3475 | 0.2175 |
No log | 25.72 | 52 | 1.9224 | 0.465 | 0.1125 | 0.1425 | 0.33 | 0.3825 | 0.2225 |
No log | 26.72 | 54 | 1.8714 | 0.4725 | 0.1125 | 0.1425 | 0.33 | 0.3875 | 0.23 |
No log | 27.72 | 56 | 1.8340 | 0.5075 | 0.1125 | 0.1425 | 0.3275 | 0.395 | 0.255 |
No log | 28.72 | 58 | 1.8065 | 0.52 | 0.115 | 0.1425 | 0.325 | 0.3975 | 0.33 |
No log | 29.72 | 60 | 1.7904 | 0.5225 | 0.115 | 0.1475 | 0.3275 | 0.41 | 0.375 |
No log | 30.72 | 62 | 1.7503 | 0.5225 | 0.115 | 0.15 | 0.3275 | 0.435 | 0.435 |
No log | 31.72 | 64 | 1.7100 | 0.53 | 0.1125 | 0.1475 | 0.34 | 0.4575 | 0.4625 |
No log | 32.72 | 66 | 1.6820 | 0.5375 | 0.1125 | 0.1475 | 0.3525 | 0.4575 | 0.495 |
No log | 33.72 | 68 | 1.6534 | 0.5475 | 0.1125 | 0.1475 | 0.3575 | 0.4725 | 0.4975 |
No log | 34.72 | 70 | 1.6230 | 0.575 | 0.1125 | 0.1475 | 0.3625 | 0.485 | 0.525 |
No log | 35.72 | 72 | 1.5960 | 0.585 | 0.115 | 0.1475 | 0.3575 | 0.5125 | 0.55 |
No log | 36.72 | 74 | 1.5728 | 0.575 | 0.11 | 0.1475 | 0.365 | 0.52 | 0.5525 |
No log | 37.72 | 76 | 1.5469 | 0.585 | 0.11 | 0.1475 | 0.37 | 0.525 | 0.5525 |
No log | 38.72 | 78 | 1.5219 | 0.5975 | 0.11 | 0.15 | 0.37 | 0.5275 | 0.56 |
No log | 39.72 | 80 | 1.5011 | 0.61 | 0.1125 | 0.15 | 0.3725 | 0.5375 | 0.565 |
No log | 40.72 | 82 | 1.4998 | 0.61 | 0.1125 | 0.155 | 0.37 | 0.53 | 0.57 |
No log | 41.72 | 84 | 1.4762 | 0.6175 | 0.1125 | 0.155 | 0.375 | 0.53 | 0.575 |
No log | 42.72 | 86 | 1.4490 | 0.6325 | 0.1125 | 0.1575 | 0.38 | 0.5375 | 0.585 |
No log | 43.72 | 88 | 1.4278 | 0.62 | 0.1175 | 0.1575 | 0.375 | 0.54 | 0.59 |
No log | 44.72 | 90 | 1.4169 | 0.63 | 0.115 | 0.1575 | 0.3775 | 0.54 | 0.585 |
No log | 45.72 | 92 | 1.4143 | 0.6225 | 0.115 | 0.1575 | 0.385 | 0.535 | 0.5925 |
No log | 46.72 | 94 | 1.3997 | 0.615 | 0.1175 | 0.1575 | 0.385 | 0.5425 | 0.6 |
No log | 47.72 | 96 | 1.3838 | 0.6225 | 0.1175 | 0.1575 | 0.385 | 0.54 | 0.6025 |
No log | 48.72 | 98 | 1.3737 | 0.625 | 0.115 | 0.16 | 0.39 | 0.545 | 0.5975 |
No log | 49.72 | 100 | 1.3668 | 0.6375 | 0.1125 | 0.16 | 0.3925 | 0.55 | 0.6025 |
No log | 50.72 | 102 | 1.3623 | 0.6375 | 0.1125 | 0.16 | 0.39 | 0.5525 | 0.6075 |
No log | 51.72 | 104 | 1.3525 | 0.635 | 0.1125 | 0.16 | 0.39 | 0.5525 | 0.6125 |
No log | 52.72 | 106 | 1.3455 | 0.6375 | 0.1125 | 0.16 | 0.395 | 0.555 | 0.61 |
No log | 53.72 | 108 | 1.3390 | 0.64 | 0.1125 | 0.16 | 0.3925 | 0.5525 | 0.6075 |
No log | 54.72 | 110 | 1.3338 | 0.645 | 0.11 | 0.16 | 0.3875 | 0.5525 | 0.605 |
No log | 55.72 | 112 | 1.3298 | 0.645 | 0.1125 | 0.16 | 0.385 | 0.5525 | 0.61 |
No log | 56.72 | 114 | 1.3262 | 0.645 | 0.1125 | 0.16 | 0.385 | 0.5525 | 0.6075 |
No log | 57.72 | 116 | 1.3241 | 0.6425 | 0.115 | 0.16 | 0.385 | 0.5525 | 0.6075 |
No log | 58.72 | 118 | 1.3237 | 0.6425 | 0.1125 | 0.1625 | 0.385 | 0.555 | 0.6075 |
No log | 59.72 | 120 | 1.3237 | 0.6425 | 0.1125 | 0.1625 | 0.385 | 0.555 | 0.6075 |
Framework versions
- Transformers 4.26.1
- Pytorch 1.13.1.post200
- Datasets 2.9.0
- Tokenizers 0.13.2