smids_1x_deit_tiny_rms_0001_fold3
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: 1.0984
- Accuracy: 0.8817
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 |
|---|---|---|---|---|
| 0.7593 | 1.0 | 75 | 0.5650 | 0.795 |
| 0.4619 | 2.0 | 150 | 0.5773 | 0.74 |
| 0.3635 | 3.0 | 225 | 0.4403 | 0.85 |
| 0.3309 | 4.0 | 300 | 0.3368 | 0.865 |
| 0.2141 | 5.0 | 375 | 0.4183 | 0.86 |
| 0.1728 | 6.0 | 450 | 0.5130 | 0.8617 |
| 0.154 | 7.0 | 525 | 0.4040 | 0.8683 |
| 0.1331 | 8.0 | 600 | 0.5264 | 0.855 |
| 0.0922 | 9.0 | 675 | 0.7609 | 0.8467 |
| 0.0818 | 10.0 | 750 | 0.6611 | 0.8733 |
| 0.1085 | 11.0 | 825 | 0.6834 | 0.8783 |
| 0.0246 | 12.0 | 900 | 0.6979 | 0.8783 |
| 0.0878 | 13.0 | 975 | 0.9032 | 0.8433 |
| 0.1089 | 14.0 | 1050 | 0.7716 | 0.86 |
| 0.0143 | 15.0 | 1125 | 0.8328 | 0.8633 |
| 0.0139 | 16.0 | 1200 | 0.8299 | 0.86 |
| 0.0102 | 17.0 | 1275 | 0.9296 | 0.865 |
| 0.1127 | 18.0 | 1350 | 0.8373 | 0.8517 |
| 0.0195 | 19.0 | 1425 | 0.9545 | 0.87 |
| 0.0133 | 20.0 | 1500 | 1.0697 | 0.8617 |
| 0.0108 | 21.0 | 1575 | 1.2111 | 0.8333 |
| 0.0129 | 22.0 | 1650 | 1.0019 | 0.8667 |
| 0.0225 | 23.0 | 1725 | 1.0409 | 0.8567 |
| 0.0251 | 24.0 | 1800 | 1.1816 | 0.8467 |
| 0.0012 | 25.0 | 1875 | 1.1276 | 0.8633 |
| 0.003 | 26.0 | 1950 | 1.2613 | 0.8517 |
| 0.0173 | 27.0 | 2025 | 1.2463 | 0.8667 |
| 0.0413 | 28.0 | 2100 | 1.0810 | 0.8667 |
| 0.0002 | 29.0 | 2175 | 1.0875 | 0.8683 |
| 0.0257 | 30.0 | 2250 | 1.0511 | 0.86 |
| 0.0 | 31.0 | 2325 | 0.9064 | 0.895 |
| 0.0 | 32.0 | 2400 | 0.9226 | 0.8933 |
| 0.0 | 33.0 | 2475 | 0.9451 | 0.8933 |
| 0.0062 | 34.0 | 2550 | 0.9369 | 0.8917 |
| 0.0083 | 35.0 | 2625 | 0.9494 | 0.885 |
| 0.0047 | 36.0 | 2700 | 0.9547 | 0.885 |
| 0.0019 | 37.0 | 2775 | 0.9760 | 0.885 |
| 0.0 | 38.0 | 2850 | 1.0147 | 0.8883 |
| 0.0 | 39.0 | 2925 | 0.9604 | 0.885 |
| 0.0 | 40.0 | 3000 | 0.9768 | 0.8883 |
| 0.0 | 41.0 | 3075 | 0.9916 | 0.8817 |
| 0.003 | 42.0 | 3150 | 1.0329 | 0.8867 |
| 0.0028 | 43.0 | 3225 | 1.1065 | 0.8733 |
| 0.0 | 44.0 | 3300 | 1.0908 | 0.8817 |
| 0.0025 | 45.0 | 3375 | 1.0917 | 0.8817 |
| 0.0025 | 46.0 | 3450 | 1.0917 | 0.8817 |
| 0.0051 | 47.0 | 3525 | 1.0933 | 0.8817 |
| 0.0 | 48.0 | 3600 | 1.0955 | 0.8817 |
| 0.0 | 49.0 | 3675 | 1.0973 | 0.8817 |
| 0.0044 | 50.0 | 3750 | 1.0984 | 0.8817 |
Framework versions
- Transformers 4.35.2
- Pytorch 2.1.0+cu118
- Datasets 2.15.0
- Tokenizers 0.15.0