outlmv3_jordyvl_rvl_cdip_100_examples_per_class_2023-12-01_txt_vis_concat_enc_1_2_3_4_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: 1.0852
- Accuracy: 0.755
- Exit 0 Accuracy: 0.06
- Exit 1 Accuracy: 0.0625
- Exit 2 Accuracy: 0.0575
- Exit 3 Accuracy: 0.065
- Exit 4 Accuracy: 0.0775
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: 2
- seed: 42
- gradient_accumulation_steps: 24
- total_train_batch_size: 96
- 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 | 8 | 2.6886 | 0.155 | 0.055 | 0.0825 | 0.0625 | 0.0625 | 0.0625 |
No log | 1.96 | 16 | 2.5967 | 0.205 | 0.05 | 0.0625 | 0.0625 | 0.0625 | 0.0625 |
No log | 2.96 | 24 | 2.4863 | 0.2275 | 0.0525 | 0.0625 | 0.0625 | 0.08 | 0.0625 |
No log | 3.96 | 32 | 2.3521 | 0.285 | 0.045 | 0.07 | 0.0725 | 0.065 | 0.065 |
No log | 4.96 | 40 | 2.2600 | 0.3025 | 0.04 | 0.0675 | 0.05 | 0.065 | 0.0675 |
No log | 5.96 | 48 | 2.1585 | 0.3425 | 0.035 | 0.0675 | 0.0675 | 0.0625 | 0.0625 |
No log | 6.96 | 56 | 2.0467 | 0.41 | 0.0375 | 0.065 | 0.065 | 0.0575 | 0.0625 |
No log | 7.96 | 64 | 1.8287 | 0.525 | 0.0375 | 0.0575 | 0.0525 | 0.0625 | 0.0625 |
No log | 8.96 | 72 | 1.6875 | 0.5775 | 0.045 | 0.065 | 0.055 | 0.065 | 0.0575 |
No log | 9.96 | 80 | 1.5657 | 0.5925 | 0.045 | 0.0575 | 0.05 | 0.065 | 0.0575 |
No log | 10.96 | 88 | 1.4216 | 0.6325 | 0.05 | 0.075 | 0.045 | 0.0625 | 0.0525 |
No log | 11.96 | 96 | 1.3001 | 0.6575 | 0.055 | 0.075 | 0.05 | 0.065 | 0.055 |
No log | 12.96 | 104 | 1.2468 | 0.67 | 0.0575 | 0.075 | 0.055 | 0.065 | 0.0625 |
No log | 13.96 | 112 | 1.1777 | 0.685 | 0.055 | 0.075 | 0.0575 | 0.06 | 0.055 |
No log | 14.96 | 120 | 1.1468 | 0.6875 | 0.055 | 0.0775 | 0.0525 | 0.06 | 0.0475 |
No log | 15.96 | 128 | 1.0561 | 0.72 | 0.055 | 0.08 | 0.0525 | 0.06 | 0.055 |
No log | 16.96 | 136 | 1.0213 | 0.7175 | 0.055 | 0.085 | 0.05 | 0.0625 | 0.055 |
No log | 17.96 | 144 | 1.0266 | 0.7125 | 0.055 | 0.085 | 0.055 | 0.06 | 0.055 |
No log | 18.96 | 152 | 0.9733 | 0.7275 | 0.0525 | 0.0875 | 0.0475 | 0.06 | 0.0625 |
No log | 19.96 | 160 | 0.9511 | 0.7475 | 0.0525 | 0.0775 | 0.055 | 0.06 | 0.065 |
No log | 20.96 | 168 | 0.9595 | 0.735 | 0.0525 | 0.0675 | 0.055 | 0.06 | 0.06 |
No log | 21.96 | 176 | 0.9803 | 0.7475 | 0.055 | 0.0675 | 0.055 | 0.06 | 0.07 |
No log | 22.96 | 184 | 0.9428 | 0.75 | 0.0575 | 0.0675 | 0.0525 | 0.06 | 0.08 |
No log | 23.96 | 192 | 0.9591 | 0.7275 | 0.0525 | 0.065 | 0.05 | 0.06 | 0.0825 |
No log | 24.96 | 200 | 0.9216 | 0.7525 | 0.06 | 0.065 | 0.055 | 0.06 | 0.0825 |
No log | 25.96 | 208 | 0.9194 | 0.7525 | 0.0575 | 0.065 | 0.0525 | 0.06 | 0.075 |
No log | 26.96 | 216 | 1.0271 | 0.7275 | 0.0575 | 0.065 | 0.0525 | 0.06 | 0.075 |
No log | 27.96 | 224 | 0.9563 | 0.77 | 0.0625 | 0.065 | 0.0525 | 0.06 | 0.0775 |
No log | 28.96 | 232 | 0.9999 | 0.7275 | 0.0625 | 0.065 | 0.055 | 0.06 | 0.0675 |
No log | 29.96 | 240 | 0.9599 | 0.76 | 0.0625 | 0.065 | 0.05 | 0.06 | 0.065 |
No log | 30.96 | 248 | 0.9884 | 0.75 | 0.0625 | 0.065 | 0.05 | 0.06 | 0.0625 |
No log | 31.96 | 256 | 1.0037 | 0.745 | 0.0625 | 0.0625 | 0.0525 | 0.06 | 0.0625 |
No log | 32.96 | 264 | 0.9848 | 0.7425 | 0.0625 | 0.0625 | 0.0525 | 0.0625 | 0.065 |
No log | 33.96 | 272 | 1.0081 | 0.7525 | 0.0625 | 0.0625 | 0.055 | 0.0625 | 0.065 |
No log | 34.96 | 280 | 1.0274 | 0.755 | 0.0575 | 0.0625 | 0.055 | 0.0625 | 0.075 |
No log | 35.96 | 288 | 1.0378 | 0.7525 | 0.0575 | 0.0625 | 0.055 | 0.0675 | 0.0725 |
No log | 36.96 | 296 | 1.0480 | 0.7525 | 0.0625 | 0.065 | 0.0525 | 0.065 | 0.07 |
No log | 37.96 | 304 | 1.0332 | 0.765 | 0.0625 | 0.065 | 0.0525 | 0.065 | 0.075 |
No log | 38.96 | 312 | 1.0222 | 0.765 | 0.06 | 0.0625 | 0.0525 | 0.065 | 0.0775 |
No log | 39.96 | 320 | 1.0709 | 0.75 | 0.06 | 0.0625 | 0.0525 | 0.065 | 0.08 |
No log | 40.96 | 328 | 1.0426 | 0.755 | 0.06 | 0.0625 | 0.0525 | 0.065 | 0.08 |
No log | 41.96 | 336 | 1.0789 | 0.74 | 0.06 | 0.0625 | 0.0525 | 0.065 | 0.08 |
No log | 42.96 | 344 | 1.0492 | 0.765 | 0.06 | 0.0625 | 0.0525 | 0.0625 | 0.0775 |
No log | 43.96 | 352 | 1.0541 | 0.7575 | 0.06 | 0.0625 | 0.0525 | 0.0625 | 0.08 |
No log | 44.96 | 360 | 1.0620 | 0.755 | 0.06 | 0.0625 | 0.0525 | 0.06 | 0.08 |
No log | 45.96 | 368 | 1.0514 | 0.7575 | 0.06 | 0.0625 | 0.055 | 0.06 | 0.0775 |
No log | 46.96 | 376 | 1.0537 | 0.755 | 0.06 | 0.0625 | 0.0525 | 0.0625 | 0.0775 |
No log | 47.96 | 384 | 1.0662 | 0.7575 | 0.06 | 0.0625 | 0.0525 | 0.0625 | 0.0775 |
No log | 48.96 | 392 | 1.0693 | 0.76 | 0.06 | 0.0625 | 0.055 | 0.0625 | 0.08 |
No log | 49.96 | 400 | 1.0775 | 0.7575 | 0.06 | 0.0625 | 0.055 | 0.0625 | 0.08 |
No log | 50.96 | 408 | 1.0863 | 0.75 | 0.0575 | 0.0625 | 0.0575 | 0.065 | 0.0825 |
No log | 51.96 | 416 | 1.0567 | 0.76 | 0.06 | 0.0625 | 0.055 | 0.0625 | 0.08 |
No log | 52.96 | 424 | 1.0605 | 0.76 | 0.06 | 0.0625 | 0.055 | 0.0625 | 0.08 |
No log | 53.96 | 432 | 1.0720 | 0.755 | 0.06 | 0.0625 | 0.0525 | 0.0625 | 0.08 |
No log | 54.96 | 440 | 1.0807 | 0.7525 | 0.06 | 0.0625 | 0.0575 | 0.0625 | 0.0775 |
No log | 55.96 | 448 | 1.0747 | 0.7575 | 0.06 | 0.0625 | 0.055 | 0.0625 | 0.0775 |
No log | 56.96 | 456 | 1.0740 | 0.755 | 0.06 | 0.0625 | 0.06 | 0.065 | 0.08 |
No log | 57.96 | 464 | 1.0813 | 0.755 | 0.06 | 0.0625 | 0.0575 | 0.065 | 0.0775 |
No log | 58.96 | 472 | 1.0852 | 0.755 | 0.06 | 0.0625 | 0.0575 | 0.065 | 0.0775 |
No log | 59.96 | 480 | 1.0852 | 0.755 | 0.06 | 0.0625 | 0.0575 | 0.065 | 0.0775 |
Framework versions
- Transformers 4.26.1
- Pytorch 1.13.1.post200
- Datasets 2.9.0
- Tokenizers 0.13.2
- Downloads last month
- 0