smids_5x_deit_small_sgd_0001_fold3
This model is a fine-tuned version of facebook/deit-small-patch16-224 on the imagefolder dataset. It achieves the following results on the evaluation set:
- Loss: 0.4989
- Accuracy: 0.835
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 |
|---|---|---|---|---|
| 1.059 | 1.0 | 375 | 1.0700 | 0.43 |
| 1.011 | 2.0 | 750 | 1.0336 | 0.4883 |
| 0.9442 | 3.0 | 1125 | 0.9972 | 0.5233 |
| 0.9044 | 4.0 | 1500 | 0.9613 | 0.5667 |
| 0.9034 | 5.0 | 1875 | 0.9273 | 0.6067 |
| 0.8397 | 6.0 | 2250 | 0.8960 | 0.6217 |
| 0.8344 | 7.0 | 2625 | 0.8666 | 0.6533 |
| 0.8102 | 8.0 | 3000 | 0.8388 | 0.66 |
| 0.7533 | 9.0 | 3375 | 0.8130 | 0.675 |
| 0.7704 | 10.0 | 3750 | 0.7889 | 0.6783 |
| 0.6922 | 11.0 | 4125 | 0.7657 | 0.695 |
| 0.7058 | 12.0 | 4500 | 0.7444 | 0.7067 |
| 0.7015 | 13.0 | 4875 | 0.7244 | 0.7183 |
| 0.7084 | 14.0 | 5250 | 0.7056 | 0.725 |
| 0.6276 | 15.0 | 5625 | 0.6882 | 0.74 |
| 0.6138 | 16.0 | 6000 | 0.6721 | 0.745 |
| 0.6401 | 17.0 | 6375 | 0.6573 | 0.7533 |
| 0.6373 | 18.0 | 6750 | 0.6430 | 0.7633 |
| 0.569 | 19.0 | 7125 | 0.6303 | 0.7633 |
| 0.5819 | 20.0 | 7500 | 0.6185 | 0.77 |
| 0.5294 | 21.0 | 7875 | 0.6077 | 0.7817 |
| 0.5473 | 22.0 | 8250 | 0.5978 | 0.7883 |
| 0.5629 | 23.0 | 8625 | 0.5888 | 0.7967 |
| 0.5783 | 24.0 | 9000 | 0.5802 | 0.8017 |
| 0.509 | 25.0 | 9375 | 0.5724 | 0.8067 |
| 0.5255 | 26.0 | 9750 | 0.5652 | 0.805 |
| 0.5612 | 27.0 | 10125 | 0.5585 | 0.81 |
| 0.5914 | 28.0 | 10500 | 0.5523 | 0.815 |
| 0.4839 | 29.0 | 10875 | 0.5467 | 0.815 |
| 0.4781 | 30.0 | 11250 | 0.5414 | 0.8167 |
| 0.5423 | 31.0 | 11625 | 0.5367 | 0.8183 |
| 0.5434 | 32.0 | 12000 | 0.5323 | 0.8183 |
| 0.5812 | 33.0 | 12375 | 0.5281 | 0.82 |
| 0.4776 | 34.0 | 12750 | 0.5244 | 0.8183 |
| 0.4385 | 35.0 | 13125 | 0.5209 | 0.8217 |
| 0.4956 | 36.0 | 13500 | 0.5178 | 0.8217 |
| 0.4746 | 37.0 | 13875 | 0.5150 | 0.8233 |
| 0.4824 | 38.0 | 14250 | 0.5124 | 0.8233 |
| 0.49 | 39.0 | 14625 | 0.5101 | 0.8217 |
| 0.4379 | 40.0 | 15000 | 0.5080 | 0.8233 |
| 0.4149 | 41.0 | 15375 | 0.5062 | 0.825 |
| 0.4917 | 42.0 | 15750 | 0.5046 | 0.8267 |
| 0.5208 | 43.0 | 16125 | 0.5031 | 0.8283 |
| 0.4676 | 44.0 | 16500 | 0.5020 | 0.8283 |
| 0.4552 | 45.0 | 16875 | 0.5009 | 0.8317 |
| 0.4563 | 46.0 | 17250 | 0.5002 | 0.8333 |
| 0.5467 | 47.0 | 17625 | 0.4996 | 0.835 |
| 0.5056 | 48.0 | 18000 | 0.4992 | 0.835 |
| 0.4817 | 49.0 | 18375 | 0.4990 | 0.835 |
| 0.4808 | 50.0 | 18750 | 0.4989 | 0.835 |
Framework versions
- Transformers 4.32.1
- Pytorch 2.1.1+cu121
- Datasets 2.12.0
- Tokenizers 0.13.2