smids_3x_deit_tiny_sgd_001_fold1
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: 0.3078
- Accuracy: 0.8731
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: 0.001
- 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 |
|---|---|---|---|---|
| 0.9352 | 1.0 | 226 | 0.9208 | 0.5376 |
| 0.6928 | 2.0 | 452 | 0.7389 | 0.6861 |
| 0.5623 | 3.0 | 678 | 0.6105 | 0.7429 |
| 0.5246 | 4.0 | 904 | 0.5485 | 0.7563 |
| 0.5426 | 5.0 | 1130 | 0.4979 | 0.7880 |
| 0.4977 | 6.0 | 1356 | 0.4581 | 0.8080 |
| 0.3766 | 7.0 | 1582 | 0.4327 | 0.8130 |
| 0.4038 | 8.0 | 1808 | 0.4167 | 0.8097 |
| 0.3541 | 9.0 | 2034 | 0.3987 | 0.8431 |
| 0.3195 | 10.0 | 2260 | 0.3857 | 0.8247 |
| 0.3215 | 11.0 | 2486 | 0.3815 | 0.8297 |
| 0.2707 | 12.0 | 2712 | 0.3604 | 0.8414 |
| 0.2756 | 13.0 | 2938 | 0.3575 | 0.8364 |
| 0.2853 | 14.0 | 3164 | 0.3492 | 0.8414 |
| 0.3202 | 15.0 | 3390 | 0.3434 | 0.8447 |
| 0.3213 | 16.0 | 3616 | 0.3398 | 0.8497 |
| 0.246 | 17.0 | 3842 | 0.3305 | 0.8581 |
| 0.2485 | 18.0 | 4068 | 0.3288 | 0.8564 |
| 0.2691 | 19.0 | 4294 | 0.3315 | 0.8598 |
| 0.2123 | 20.0 | 4520 | 0.3213 | 0.8648 |
| 0.2607 | 21.0 | 4746 | 0.3252 | 0.8564 |
| 0.2646 | 22.0 | 4972 | 0.3186 | 0.8664 |
| 0.2851 | 23.0 | 5198 | 0.3202 | 0.8631 |
| 0.2373 | 24.0 | 5424 | 0.3144 | 0.8748 |
| 0.1908 | 25.0 | 5650 | 0.3143 | 0.8698 |
| 0.2924 | 26.0 | 5876 | 0.3120 | 0.8698 |
| 0.1662 | 27.0 | 6102 | 0.3113 | 0.8748 |
| 0.2215 | 28.0 | 6328 | 0.3120 | 0.8681 |
| 0.1838 | 29.0 | 6554 | 0.3136 | 0.8698 |
| 0.2131 | 30.0 | 6780 | 0.3140 | 0.8731 |
| 0.2074 | 31.0 | 7006 | 0.3100 | 0.8715 |
| 0.194 | 32.0 | 7232 | 0.3083 | 0.8748 |
| 0.1635 | 33.0 | 7458 | 0.3091 | 0.8748 |
| 0.1521 | 34.0 | 7684 | 0.3083 | 0.8748 |
| 0.2333 | 35.0 | 7910 | 0.3078 | 0.8748 |
| 0.1942 | 36.0 | 8136 | 0.3076 | 0.8731 |
| 0.242 | 37.0 | 8362 | 0.3062 | 0.8748 |
| 0.2131 | 38.0 | 8588 | 0.3090 | 0.8748 |
| 0.2044 | 39.0 | 8814 | 0.3079 | 0.8748 |
| 0.1565 | 40.0 | 9040 | 0.3082 | 0.8731 |
| 0.1709 | 41.0 | 9266 | 0.3089 | 0.8748 |
| 0.2023 | 42.0 | 9492 | 0.3080 | 0.8748 |
| 0.2299 | 43.0 | 9718 | 0.3077 | 0.8731 |
| 0.1365 | 44.0 | 9944 | 0.3081 | 0.8765 |
| 0.1955 | 45.0 | 10170 | 0.3078 | 0.8748 |
| 0.2025 | 46.0 | 10396 | 0.3089 | 0.8781 |
| 0.1982 | 47.0 | 10622 | 0.3076 | 0.8731 |
| 0.1881 | 48.0 | 10848 | 0.3078 | 0.8731 |
| 0.1389 | 49.0 | 11074 | 0.3077 | 0.8731 |
| 0.1646 | 50.0 | 11300 | 0.3078 | 0.8731 |
Framework versions
- Transformers 4.32.1
- Pytorch 2.1.1+cu121
- Datasets 2.12.0
- Tokenizers 0.13.2