smids_5x_deit_tiny_sgd_0001_fold2
This model is a fine-tuned version of facebook/deit-tiny-patch16-224 on the imagefolder dataset. It achieves the following results on the evaluation set:
- Loss: 0.4980
- Accuracy: 0.8003
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: 0.0001
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- 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 |
|---|---|---|---|---|
| 1.1772 | 1.0 | 375 | 1.1487 | 0.3760 |
| 1.0767 | 2.0 | 750 | 1.0817 | 0.4276 |
| 0.9966 | 3.0 | 1125 | 1.0309 | 0.4742 |
| 0.9479 | 4.0 | 1500 | 0.9881 | 0.5158 |
| 0.9443 | 5.0 | 1875 | 0.9501 | 0.5374 |
| 0.9172 | 6.0 | 2250 | 0.9138 | 0.5574 |
| 0.8419 | 7.0 | 2625 | 0.8792 | 0.5857 |
| 0.8745 | 8.0 | 3000 | 0.8452 | 0.6023 |
| 0.8037 | 9.0 | 3375 | 0.8135 | 0.6273 |
| 0.7725 | 10.0 | 3750 | 0.7849 | 0.6556 |
| 0.73 | 11.0 | 4125 | 0.7582 | 0.6705 |
| 0.7472 | 12.0 | 4500 | 0.7325 | 0.6822 |
| 0.6618 | 13.0 | 4875 | 0.7082 | 0.7105 |
| 0.6872 | 14.0 | 5250 | 0.6883 | 0.7255 |
| 0.6348 | 15.0 | 5625 | 0.6702 | 0.7304 |
| 0.5716 | 16.0 | 6000 | 0.6521 | 0.7388 |
| 0.6306 | 17.0 | 6375 | 0.6373 | 0.7454 |
| 0.604 | 18.0 | 6750 | 0.6231 | 0.7537 |
| 0.5325 | 19.0 | 7125 | 0.6109 | 0.7604 |
| 0.602 | 20.0 | 7500 | 0.6001 | 0.7604 |
| 0.5143 | 21.0 | 7875 | 0.5907 | 0.7604 |
| 0.5394 | 22.0 | 8250 | 0.5814 | 0.7671 |
| 0.5319 | 23.0 | 8625 | 0.5730 | 0.7671 |
| 0.5182 | 24.0 | 9000 | 0.5659 | 0.7704 |
| 0.5353 | 25.0 | 9375 | 0.5592 | 0.7720 |
| 0.5323 | 26.0 | 9750 | 0.5527 | 0.7787 |
| 0.5286 | 27.0 | 10125 | 0.5472 | 0.7787 |
| 0.4649 | 28.0 | 10500 | 0.5420 | 0.7820 |
| 0.5396 | 29.0 | 10875 | 0.5373 | 0.7837 |
| 0.4863 | 30.0 | 11250 | 0.5330 | 0.7837 |
| 0.4735 | 31.0 | 11625 | 0.5292 | 0.7837 |
| 0.5595 | 32.0 | 12000 | 0.5255 | 0.7887 |
| 0.4938 | 33.0 | 12375 | 0.5219 | 0.7920 |
| 0.468 | 34.0 | 12750 | 0.5190 | 0.7970 |
| 0.4879 | 35.0 | 13125 | 0.5162 | 0.7970 |
| 0.4592 | 36.0 | 13500 | 0.5136 | 0.7970 |
| 0.4625 | 37.0 | 13875 | 0.5113 | 0.7953 |
| 0.4476 | 38.0 | 14250 | 0.5091 | 0.7970 |
| 0.4681 | 39.0 | 14625 | 0.5073 | 0.8003 |
| 0.4644 | 40.0 | 15000 | 0.5056 | 0.8003 |
| 0.4624 | 41.0 | 15375 | 0.5040 | 0.8003 |
| 0.4244 | 42.0 | 15750 | 0.5026 | 0.8003 |
| 0.457 | 43.0 | 16125 | 0.5015 | 0.8003 |
| 0.4802 | 44.0 | 16500 | 0.5005 | 0.8003 |
| 0.4285 | 45.0 | 16875 | 0.4996 | 0.8003 |
| 0.4665 | 46.0 | 17250 | 0.4990 | 0.8003 |
| 0.4627 | 47.0 | 17625 | 0.4985 | 0.8003 |
| 0.4554 | 48.0 | 18000 | 0.4982 | 0.8003 |
| 0.4888 | 49.0 | 18375 | 0.4980 | 0.8003 |
| 0.4643 | 50.0 | 18750 | 0.4980 | 0.8003 |
Framework versions
- Transformers 4.32.1
- Pytorch 2.1.1+cu121
- Datasets 2.12.0
- Tokenizers 0.13.2