pairs-classifier-20x-large-516-line100
This model is a fine-tuned version of facebook/dinov2-large-imagenet1k-1-layer on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 1.1753
- Accuracy: 0.6791
- F1: 0.6790
- Recall: 0.7219
- Precision: 0.7241
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: 5e-05
- train_batch_size: 64
- eval_batch_size: 64
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 256
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 50
- mixed_precision_training: Native AMP
Training results
Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 | Recall | Precision |
---|---|---|---|---|---|---|---|
0.7285 | 0.95 | 13 | 0.6921 | 0.6279 | 0.3857 | 0.3140 | 0.5 |
0.6882 | 1.96 | 27 | 0.6537 | 0.6372 | 0.5580 | 0.5947 | 0.5660 |
0.5771 | 2.98 | 41 | 0.6030 | 0.7070 | 0.7039 | 0.7141 | 0.7285 |
0.5086 | 4.0 | 55 | 0.5027 | 0.7395 | 0.7365 | 0.7453 | 0.7620 |
0.4418 | 4.95 | 68 | 0.6439 | 0.7023 | 0.7022 | 0.7652 | 0.7579 |
0.3432 | 5.96 | 82 | 0.6873 | 0.7116 | 0.7115 | 0.7755 | 0.7678 |
0.3346 | 6.98 | 96 | 0.4600 | 0.7721 | 0.7579 | 0.7563 | 0.7600 |
0.2755 | 8.0 | 110 | 0.8594 | 0.6884 | 0.6881 | 0.7528 | 0.7442 |
0.2676 | 8.95 | 123 | 0.9163 | 0.6884 | 0.6879 | 0.7588 | 0.7468 |
0.2658 | 9.96 | 137 | 0.5970 | 0.7442 | 0.7424 | 0.7575 | 0.7734 |
0.24 | 10.98 | 151 | 0.5749 | 0.7488 | 0.7454 | 0.7519 | 0.7694 |
0.234 | 12.0 | 165 | 0.8778 | 0.6977 | 0.6976 | 0.7517 | 0.7491 |
0.2154 | 12.95 | 178 | 0.8210 | 0.7023 | 0.7023 | 0.7490 | 0.7502 |
0.203 | 13.96 | 192 | 0.7305 | 0.7163 | 0.7154 | 0.7397 | 0.7512 |
0.1765 | 14.98 | 206 | 0.8578 | 0.6977 | 0.6976 | 0.7416 | 0.7440 |
0.1607 | 16.0 | 220 | 0.6878 | 0.7581 | 0.7543 | 0.7587 | 0.7769 |
0.1783 | 16.95 | 233 | 0.6481 | 0.7814 | 0.7764 | 0.7765 | 0.7954 |
0.1767 | 17.96 | 247 | 0.9805 | 0.6791 | 0.6790 | 0.7174 | 0.7215 |
0.1596 | 18.98 | 261 | 1.0490 | 0.6977 | 0.6976 | 0.7517 | 0.7491 |
0.152 | 20.0 | 275 | 1.1249 | 0.6744 | 0.6741 | 0.7402 | 0.7306 |
0.1807 | 20.95 | 288 | 0.8215 | 0.7302 | 0.7281 | 0.7419 | 0.7572 |
0.1441 | 21.96 | 302 | 1.2401 | 0.6791 | 0.6788 | 0.7425 | 0.7343 |
0.1494 | 22.98 | 316 | 0.9281 | 0.7023 | 0.7015 | 0.7273 | 0.7375 |
0.1303 | 24.0 | 330 | 0.8422 | 0.7256 | 0.7236 | 0.7387 | 0.7535 |
0.1324 | 24.95 | 343 | 0.9261 | 0.7070 | 0.7061 | 0.7302 | 0.7412 |
0.1214 | 25.96 | 357 | 1.1154 | 0.6791 | 0.6791 | 0.7266 | 0.7266 |
0.1184 | 26.98 | 371 | 1.2263 | 0.6698 | 0.6697 | 0.7267 | 0.7218 |
0.14 | 28.0 | 385 | 0.9725 | 0.7163 | 0.7154 | 0.7397 | 0.7512 |
0.107 | 28.95 | 398 | 0.9908 | 0.7163 | 0.7154 | 0.7397 | 0.7512 |
0.1082 | 29.96 | 412 | 1.1253 | 0.6698 | 0.6697 | 0.7216 | 0.7192 |
0.1477 | 30.98 | 426 | 0.7786 | 0.7581 | 0.7537 | 0.7564 | 0.7743 |
0.1296 | 32.0 | 440 | 1.5284 | 0.6558 | 0.6546 | 0.7377 | 0.7183 |
0.1122 | 32.95 | 453 | 1.0805 | 0.7070 | 0.7063 | 0.734 | 0.7437 |
0.0994 | 33.96 | 467 | 1.2907 | 0.6651 | 0.6651 | 0.7190 | 0.7155 |
0.1046 | 34.98 | 481 | 0.8528 | 0.7488 | 0.7454 | 0.7519 | 0.7694 |
0.0942 | 36.0 | 495 | 1.5078 | 0.6651 | 0.6645 | 0.7357 | 0.7231 |
0.1089 | 36.95 | 508 | 0.9448 | 0.7442 | 0.7415 | 0.7514 | 0.7683 |
0.096 | 37.96 | 522 | 1.2000 | 0.6651 | 0.6651 | 0.7094 | 0.7104 |
0.0903 | 38.98 | 536 | 0.9947 | 0.7163 | 0.7147 | 0.7326 | 0.7461 |
0.079 | 40.0 | 550 | 1.0776 | 0.7116 | 0.7106 | 0.7331 | 0.7449 |
0.0787 | 40.95 | 563 | 1.1923 | 0.6744 | 0.6744 | 0.7193 | 0.7204 |
0.0882 | 41.96 | 577 | 1.1516 | 0.6791 | 0.6788 | 0.7132 | 0.7190 |
0.0937 | 42.98 | 591 | 1.1568 | 0.6884 | 0.6881 | 0.7229 | 0.7289 |
0.0837 | 44.0 | 605 | 1.1792 | 0.6791 | 0.6790 | 0.7219 | 0.7241 |
0.0838 | 44.95 | 618 | 1.1944 | 0.6791 | 0.6790 | 0.7219 | 0.7241 |
0.085 | 45.96 | 632 | 1.0771 | 0.7163 | 0.7154 | 0.7397 | 0.7512 |
0.0781 | 46.98 | 646 | 1.1780 | 0.6791 | 0.6790 | 0.7219 | 0.7241 |
0.0828 | 47.27 | 650 | 1.1753 | 0.6791 | 0.6790 | 0.7219 | 0.7241 |
Framework versions
- Transformers 4.37.2
- Pytorch 2.2.0+cu121
- Datasets 2.16.1
- Tokenizers 0.15.1
- Downloads last month
- 1