<!-- 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_2023-07-05_ent_g75
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: 1.2759
- Accuracy: 0.6825
- Exit 0 Accuracy: 0.11
- Exit 1 Accuracy: 0.155
- Exit 2 Accuracy: 0.345
- Exit 3 Accuracy: 0.425
- Exit 4 Accuracy: 0.5225
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 | 2.7601 | 0.1075 | 0.0825 | 0.0675 | 0.1025 | 0.0625 | 0.0625 |
No log | 1.72 | 4 | 2.7328 | 0.1125 | 0.07 | 0.065 | 0.1225 | 0.0625 | 0.0625 |
No log | 2.72 | 6 | 2.6968 | 0.13 | 0.075 | 0.06 | 0.1325 | 0.0625 | 0.0625 |
No log | 3.72 | 8 | 2.6594 | 0.18 | 0.075 | 0.06 | 0.1175 | 0.0625 | 0.0625 |
No log | 4.72 | 10 | 2.6206 | 0.1925 | 0.085 | 0.0575 | 0.11 | 0.0625 | 0.0625 |
No log | 5.72 | 12 | 2.5710 | 0.2125 | 0.09 | 0.08 | 0.115 | 0.0625 | 0.0625 |
No log | 6.72 | 14 | 2.5334 | 0.2275 | 0.095 | 0.08 | 0.12 | 0.0575 | 0.0625 |
No log | 7.72 | 16 | 2.5094 | 0.245 | 0.095 | 0.095 | 0.135 | 0.0725 | 0.0625 |
No log | 8.72 | 18 | 2.4631 | 0.2825 | 0.095 | 0.0975 | 0.17 | 0.0925 | 0.065 |
No log | 9.72 | 20 | 2.4152 | 0.3025 | 0.1 | 0.1275 | 0.205 | 0.1075 | 0.0625 |
No log | 10.72 | 22 | 2.3737 | 0.325 | 0.1075 | 0.1225 | 0.24 | 0.12 | 0.065 |
No log | 11.72 | 24 | 2.3302 | 0.3175 | 0.1125 | 0.1175 | 0.2375 | 0.1475 | 0.0675 |
No log | 12.72 | 26 | 2.2746 | 0.34 | 0.1125 | 0.125 | 0.255 | 0.155 | 0.095 |
No log | 13.72 | 28 | 2.2527 | 0.35 | 0.1125 | 0.125 | 0.2625 | 0.175 | 0.095 |
No log | 14.72 | 30 | 2.2101 | 0.3425 | 0.1075 | 0.13 | 0.27 | 0.2125 | 0.095 |
No log | 15.72 | 32 | 2.1811 | 0.355 | 0.1075 | 0.14 | 0.29 | 0.24 | 0.095 |
No log | 16.72 | 34 | 2.1368 | 0.38 | 0.105 | 0.145 | 0.305 | 0.245 | 0.0925 |
No log | 17.72 | 36 | 2.0855 | 0.395 | 0.1075 | 0.145 | 0.3175 | 0.2475 | 0.095 |
No log | 18.72 | 38 | 2.0559 | 0.4 | 0.1125 | 0.145 | 0.305 | 0.255 | 0.1025 |
No log | 19.72 | 40 | 2.0277 | 0.41 | 0.115 | 0.145 | 0.295 | 0.28 | 0.105 |
No log | 20.72 | 42 | 1.9746 | 0.445 | 0.12 | 0.145 | 0.28 | 0.2875 | 0.1025 |
No log | 21.72 | 44 | 1.9346 | 0.4525 | 0.12 | 0.145 | 0.265 | 0.25 | 0.1025 |
No log | 22.72 | 46 | 1.8926 | 0.4925 | 0.12 | 0.145 | 0.255 | 0.25 | 0.11 |
No log | 23.72 | 48 | 1.8581 | 0.5025 | 0.115 | 0.1475 | 0.26 | 0.3 | 0.1075 |
No log | 24.72 | 50 | 1.8403 | 0.4975 | 0.11 | 0.1475 | 0.2725 | 0.3325 | 0.11 |
No log | 25.72 | 52 | 1.8162 | 0.5 | 0.1125 | 0.1475 | 0.2875 | 0.3575 | 0.1125 |
No log | 26.72 | 54 | 1.7562 | 0.5475 | 0.115 | 0.1475 | 0.295 | 0.3575 | 0.115 |
No log | 27.72 | 56 | 1.7205 | 0.5725 | 0.1175 | 0.15 | 0.295 | 0.37 | 0.115 |
No log | 28.72 | 58 | 1.7041 | 0.555 | 0.1175 | 0.15 | 0.295 | 0.36 | 0.11 |
No log | 29.72 | 60 | 1.7018 | 0.5525 | 0.12 | 0.15 | 0.305 | 0.36 | 0.1125 |
No log | 30.72 | 62 | 1.6532 | 0.58 | 0.12 | 0.15 | 0.3025 | 0.3725 | 0.1125 |
No log | 31.72 | 64 | 1.6218 | 0.58 | 0.12 | 0.15 | 0.3125 | 0.3725 | 0.1175 |
No log | 32.72 | 66 | 1.5888 | 0.59 | 0.115 | 0.1475 | 0.32 | 0.38 | 0.1325 |
No log | 33.72 | 68 | 1.5778 | 0.6 | 0.115 | 0.1475 | 0.315 | 0.3875 | 0.1425 |
No log | 34.72 | 70 | 1.5500 | 0.59 | 0.1225 | 0.15 | 0.315 | 0.3875 | 0.155 |
No log | 35.72 | 72 | 1.5216 | 0.61 | 0.13 | 0.15 | 0.31 | 0.3875 | 0.17 |
No log | 36.72 | 74 | 1.5024 | 0.6175 | 0.1275 | 0.15 | 0.3075 | 0.4125 | 0.1675 |
No log | 37.72 | 76 | 1.4787 | 0.615 | 0.12 | 0.1525 | 0.32 | 0.4025 | 0.165 |
No log | 38.72 | 78 | 1.4635 | 0.6175 | 0.1175 | 0.1525 | 0.325 | 0.4125 | 0.1625 |
No log | 39.72 | 80 | 1.4455 | 0.6225 | 0.12 | 0.155 | 0.3225 | 0.4225 | 0.165 |
No log | 40.72 | 82 | 1.4304 | 0.625 | 0.12 | 0.155 | 0.33 | 0.425 | 0.1675 |
No log | 41.72 | 84 | 1.4170 | 0.6425 | 0.1175 | 0.155 | 0.3325 | 0.425 | 0.21 |
No log | 42.72 | 86 | 1.4052 | 0.64 | 0.1175 | 0.155 | 0.335 | 0.4275 | 0.245 |
No log | 43.72 | 88 | 1.3965 | 0.6425 | 0.1125 | 0.155 | 0.34 | 0.4125 | 0.2775 |
No log | 44.72 | 90 | 1.3766 | 0.645 | 0.1125 | 0.155 | 0.3425 | 0.4075 | 0.315 |
No log | 45.72 | 92 | 1.3611 | 0.6575 | 0.11 | 0.155 | 0.345 | 0.41 | 0.33 |
No log | 46.72 | 94 | 1.3513 | 0.6575 | 0.11 | 0.155 | 0.3425 | 0.4175 | 0.34 |
No log | 47.72 | 96 | 1.3520 | 0.665 | 0.11 | 0.155 | 0.3425 | 0.4275 | 0.36 |
No log | 48.72 | 98 | 1.3373 | 0.67 | 0.11 | 0.155 | 0.3425 | 0.425 | 0.3875 |
No log | 49.72 | 100 | 1.3213 | 0.6775 | 0.11 | 0.155 | 0.3425 | 0.4175 | 0.405 |
No log | 50.72 | 102 | 1.3124 | 0.6825 | 0.11 | 0.155 | 0.3425 | 0.41 | 0.445 |
No log | 51.72 | 104 | 1.3080 | 0.68 | 0.1075 | 0.155 | 0.34 | 0.41 | 0.455 |
No log | 52.72 | 106 | 1.3037 | 0.675 | 0.105 | 0.1575 | 0.3425 | 0.4175 | 0.4775 |
No log | 53.72 | 108 | 1.2987 | 0.6825 | 0.11 | 0.1575 | 0.345 | 0.425 | 0.4875 |
No log | 54.72 | 110 | 1.2943 | 0.6775 | 0.1075 | 0.1575 | 0.3475 | 0.425 | 0.5025 |
No log | 55.72 | 112 | 1.2889 | 0.68 | 0.1075 | 0.1575 | 0.3475 | 0.425 | 0.51 |
No log | 56.72 | 114 | 1.2829 | 0.68 | 0.1075 | 0.16 | 0.3475 | 0.4225 | 0.5175 |
No log | 57.72 | 116 | 1.2793 | 0.68 | 0.1075 | 0.155 | 0.3475 | 0.4225 | 0.5225 |
No log | 58.72 | 118 | 1.2769 | 0.68 | 0.11 | 0.155 | 0.345 | 0.4225 | 0.5225 |
No log | 59.72 | 120 | 1.2759 | 0.6825 | 0.11 | 0.155 | 0.345 | 0.425 | 0.5225 |
Framework versions
- Transformers 4.26.1
- Pytorch 1.13.1.post200
- Datasets 2.9.0
- Tokenizers 0.13.2