convnextv2-tiny-1k-224-finetuned-eurosat-50
This model is a fine-tuned version of facebook/convnextv2-tiny-1k-224 on the imagefolder dataset. It achieves the following results on the evaluation set:
- Loss: 1.2472
- Accuracy: 0.7763
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: 16
- eval_batch_size: 16
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 64
- 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
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 1.9434 | 0.97 | 18 | 1.8549 | 0.2847 |
| 1.7722 | 2.0 | 37 | 1.6757 | 0.3661 |
| 1.5502 | 2.97 | 55 | 1.4652 | 0.4339 |
| 1.2595 | 4.0 | 74 | 1.1916 | 0.6068 |
| 0.9304 | 4.97 | 92 | 1.0282 | 0.6576 |
| 0.7333 | 6.0 | 111 | 0.8574 | 0.7051 |
| 0.6015 | 6.97 | 129 | 0.8427 | 0.6983 |
| 0.4617 | 8.0 | 148 | 0.7682 | 0.7458 |
| 0.3162 | 8.97 | 166 | 0.7453 | 0.7559 |
| 0.2249 | 10.0 | 185 | 0.7475 | 0.7661 |
| 0.1667 | 10.97 | 203 | 0.7677 | 0.7492 |
| 0.091 | 12.0 | 222 | 1.0114 | 0.7220 |
| 0.0783 | 12.97 | 240 | 1.0206 | 0.7186 |
| 0.0613 | 14.0 | 259 | 0.8466 | 0.7492 |
| 0.0703 | 14.97 | 277 | 1.1067 | 0.7119 |
| 0.0335 | 16.0 | 296 | 1.0117 | 0.7390 |
| 0.0171 | 16.97 | 314 | 0.9367 | 0.7525 |
| 0.0253 | 18.0 | 333 | 1.3196 | 0.7153 |
| 0.0201 | 18.97 | 351 | 1.0530 | 0.7525 |
| 0.0041 | 20.0 | 370 | 1.0523 | 0.7729 |
| 0.0154 | 20.97 | 388 | 1.1311 | 0.7661 |
| 0.0025 | 22.0 | 407 | 1.1477 | 0.7729 |
| 0.0036 | 22.97 | 425 | 1.1309 | 0.7627 |
| 0.002 | 24.0 | 444 | 1.1399 | 0.7729 |
| 0.0014 | 24.97 | 462 | 1.1543 | 0.7797 |
| 0.0011 | 26.0 | 481 | 1.1799 | 0.7763 |
| 0.0011 | 26.97 | 499 | 1.1579 | 0.7661 |
| 0.0009 | 28.0 | 518 | 1.1907 | 0.7627 |
| 0.0009 | 28.97 | 536 | 1.1878 | 0.7661 |
| 0.0008 | 30.0 | 555 | 1.1986 | 0.7661 |
| 0.0008 | 30.97 | 573 | 1.2051 | 0.7661 |
| 0.0007 | 32.0 | 592 | 1.2073 | 0.7661 |
| 0.0007 | 32.97 | 610 | 1.2156 | 0.7661 |
| 0.0007 | 34.0 | 629 | 1.2218 | 0.7627 |
| 0.0007 | 34.97 | 647 | 1.2173 | 0.7661 |
| 0.0006 | 36.0 | 666 | 1.2217 | 0.7729 |
| 0.0006 | 36.97 | 684 | 1.2272 | 0.7695 |
| 0.0006 | 38.0 | 703 | 1.2261 | 0.7763 |
| 0.0006 | 38.97 | 721 | 1.2305 | 0.7763 |
| 0.0006 | 40.0 | 740 | 1.2325 | 0.7763 |
| 0.0005 | 40.97 | 758 | 1.2362 | 0.7763 |
| 0.0005 | 42.0 | 777 | 1.2409 | 0.7763 |
| 0.0005 | 42.97 | 795 | 1.2422 | 0.7763 |
| 0.0005 | 44.0 | 814 | 1.2429 | 0.7729 |
| 0.0005 | 44.97 | 832 | 1.2434 | 0.7763 |
| 0.0005 | 46.0 | 851 | 1.2458 | 0.7763 |
| 0.0005 | 46.97 | 869 | 1.2468 | 0.7763 |
| 0.0005 | 48.0 | 888 | 1.2471 | 0.7763 |
| 0.0005 | 48.65 | 900 | 1.2472 | 0.7763 |
Framework versions
- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3