metadata
base_model: google/vit-base-patch16-224-in21k
library_name: peft
license: apache-2.0
metrics:
- accuracy
tags:
- generated_from_trainer
model-index:
- name: vit-base-patch16-224-in21k-lora
results: []
vit-base-patch16-224-in21k-lora
This model is a fine-tuned version of google/vit-base-patch16-224-in21k on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.2697
- Accuracy: 0.9228
- Pca Pca Loss: 1.3062
- Pca Pca Accuracy: 0.6503
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: 0.002
- train_batch_size: 128
- eval_batch_size: 128
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 512
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 50
- mixed_precision_training: Native AMP
Training results
Training Loss | Epoch | Step | Validation Loss | Accuracy | Pca Loss | Pca Accuracy |
---|---|---|---|---|---|---|
0.94 | 0.9923 | 97 | 0.4913 | 0.8846 | 1.2038 | 0.7736 |
0.817 | 1.9949 | 195 | 0.3755 | 0.8983 | 0.9604 | 0.795 |
0.8026 | 2.9974 | 293 | 0.3390 | 0.9051 | 0.9770 | 0.7767 |
0.7858 | 4.0 | 391 | 0.3173 | 0.908 | 0.9143 | 0.7867 |
0.6296 | 4.9923 | 488 | 0.3041 | 0.9123 | 0.9831 | 0.7605 |
0.8307 | 5.9949 | 586 | 0.2981 | 0.9097 | 0.9779 | 0.7575 |
0.7709 | 6.9974 | 684 | 0.2921 | 0.9141 | 0.9452 | 0.7604 |
0.658 | 8.0 | 782 | 0.2853 | 0.9166 | 1.0252 | 0.7402 |
0.6807 | 8.9923 | 879 | 0.2803 | 0.9179 | 1.0353 | 0.734 |
0.6216 | 9.9949 | 977 | 0.2814 | 0.9159 | 1.0342 | 0.7342 |
0.7122 | 10.9974 | 1075 | 0.2810 | 0.9179 | 1.1119 | 0.7147 |
0.5949 | 12.0 | 1173 | 0.2786 | 0.9171 | 1.1367 | 0.7071 |
0.6387 | 12.9923 | 1270 | 0.2773 | 0.9185 | 1.1188 | 0.7103 |
0.5631 | 13.9949 | 1368 | 0.2769 | 0.9198 | 1.0743 | 0.7207 |
0.5733 | 14.9974 | 1466 | 0.2763 | 0.919 | 1.1199 | 0.7109 |
0.576 | 16.0 | 1564 | 0.2719 | 0.9214 | 1.1115 | 0.7105 |
0.5544 | 16.9923 | 1661 | 0.2712 | 0.9213 | 1.0589 | 0.724 |
0.503 | 17.9949 | 1759 | 0.2718 | 0.9222 | 1.0503 | 0.7269 |
0.4921 | 18.9974 | 1857 | 0.2755 | 0.9205 | 1.0790 | 0.717 |
0.4738 | 20.0 | 1955 | 0.2707 | 0.9213 | 1.1020 | 0.7124 |
0.4823 | 20.9923 | 2052 | 0.2700 | 0.9215 | 1.2160 | 0.6829 |
0.5269 | 21.9949 | 2150 | 0.2709 | 0.9202 | 1.2518 | 0.6735 |
0.5386 | 22.9974 | 2248 | 0.2700 | 0.9198 | 1.2396 | 0.6723 |
0.5236 | 24.0 | 2346 | 0.2710 | 0.9206 | 1.2457 | 0.6728 |
0.4937 | 24.9923 | 2443 | 0.2701 | 0.9208 | 1.1680 | 0.6898 |
0.585 | 25.9949 | 2541 | 0.2707 | 0.9188 | 1.2159 | 0.6769 |
0.5391 | 26.9974 | 2639 | 0.2737 | 0.9199 | 1.2199 | 0.6747 |
0.4635 | 28.0 | 2737 | 0.2710 | 0.9186 | 1.2106 | 0.68 |
0.538 | 28.9923 | 2834 | 0.2698 | 0.9223 | 1.2144 | 0.6782 |
0.5182 | 29.9949 | 2932 | 0.2706 | 0.9219 | 1.2069 | 0.6808 |
0.4368 | 30.9974 | 3030 | 0.2715 | 0.921 | 1.2384 | 0.6728 |
0.5249 | 32.0 | 3128 | 0.2691 | 0.9202 | 1.2571 | 0.6685 |
0.5122 | 32.9923 | 3225 | 0.2710 | 0.9213 | 1.2628 | 0.6653 |
0.553 | 33.9949 | 3323 | 0.2734 | 0.9209 | 1.2588 | 0.6656 |
0.4843 | 34.9974 | 3421 | 0.2702 | 0.9217 | 1.2575 | 0.6667 |
0.5083 | 36.0 | 3519 | 0.2710 | 0.923 | 1.2574 | 0.6655 |
0.4537 | 36.9923 | 3616 | 0.2701 | 0.9218 | 1.2657 | 0.6635 |
0.485 | 37.9949 | 3714 | 0.2708 | 0.923 | 1.2852 | 0.6579 |
0.4307 | 38.9974 | 3812 | 0.2735 | 0.9209 | 1.2672 | 0.6611 |
0.4878 | 40.0 | 3910 | 0.2720 | 0.922 | 1.2981 | 0.652 |
0.4936 | 40.9923 | 4007 | 0.2717 | 0.9213 | 1.3003 | 0.6531 |
0.4256 | 41.9949 | 4105 | 0.2720 | 0.9209 | 1.2996 | 0.6525 |
0.4439 | 42.9974 | 4203 | 0.2709 | 0.9226 | 1.2975 | 0.6537 |
0.4468 | 44.0 | 4301 | 0.2703 | 0.9224 | 1.2981 | 0.6533 |
0.4269 | 44.9923 | 4398 | 0.2701 | 0.9222 | 1.3000 | 0.6528 |
0.4386 | 45.9949 | 4496 | 0.2696 | 0.9223 | 1.3100 | 0.6503 |
0.4434 | 46.9974 | 4594 | 0.2699 | 0.9225 | 1.3121 | 0.6493 |
0.473 | 48.0 | 4692 | 0.2698 | 0.923 | 1.3071 | 0.6502 |
0.4997 | 48.9923 | 4789 | 0.2698 | 0.9227 | 1.3061 | 0.6507 |
0.3989 | 49.6164 | 4850 | 0.2697 | 0.9228 | 1.3062 | 0.6503 |
Framework versions
- PEFT 0.13.0
- Transformers 4.45.1
- Pytorch 2.4.1+cu121
- Datasets 3.0.1
- Tokenizers 0.20.0