smids_10x_deit_tiny_sgd_00001_fold4
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: 0.9601
- Accuracy: 0.51
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: 1e-05
- 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 |
|---|---|---|---|---|
| 1.4156 | 1.0 | 750 | 1.2978 | 0.34 |
| 1.3315 | 2.0 | 1500 | 1.2425 | 0.3483 |
| 1.2993 | 3.0 | 2250 | 1.2021 | 0.37 |
| 1.2642 | 4.0 | 3000 | 1.1733 | 0.3717 |
| 1.1084 | 5.0 | 3750 | 1.1526 | 0.375 |
| 1.1915 | 6.0 | 4500 | 1.1373 | 0.3733 |
| 1.1121 | 7.0 | 5250 | 1.1248 | 0.3817 |
| 1.1023 | 8.0 | 6000 | 1.1144 | 0.39 |
| 1.0611 | 9.0 | 6750 | 1.1051 | 0.3867 |
| 1.0698 | 10.0 | 7500 | 1.0965 | 0.39 |
| 1.0512 | 11.0 | 8250 | 1.0884 | 0.4017 |
| 1.0962 | 12.0 | 9000 | 1.0808 | 0.405 |
| 1.0873 | 13.0 | 9750 | 1.0735 | 0.4117 |
| 1.0536 | 14.0 | 10500 | 1.0664 | 0.4183 |
| 1.0525 | 15.0 | 11250 | 1.0596 | 0.4283 |
| 1.026 | 16.0 | 12000 | 1.0532 | 0.4317 |
| 1.0131 | 17.0 | 12750 | 1.0470 | 0.44 |
| 0.9786 | 18.0 | 13500 | 1.0410 | 0.4433 |
| 0.9869 | 19.0 | 14250 | 1.0353 | 0.4467 |
| 0.9996 | 20.0 | 15000 | 1.0299 | 0.4517 |
| 1.0078 | 21.0 | 15750 | 1.0247 | 0.4533 |
| 0.9709 | 22.0 | 16500 | 1.0197 | 0.4617 |
| 1.009 | 23.0 | 17250 | 1.0149 | 0.4633 |
| 1.0068 | 24.0 | 18000 | 1.0104 | 0.4633 |
| 0.9737 | 25.0 | 18750 | 1.0061 | 0.47 |
| 0.9634 | 26.0 | 19500 | 1.0021 | 0.4767 |
| 0.9648 | 27.0 | 20250 | 0.9982 | 0.4783 |
| 0.931 | 28.0 | 21000 | 0.9946 | 0.485 |
| 0.993 | 29.0 | 21750 | 0.9911 | 0.4867 |
| 0.9852 | 30.0 | 22500 | 0.9879 | 0.49 |
| 0.9579 | 31.0 | 23250 | 0.9848 | 0.49 |
| 0.9747 | 32.0 | 24000 | 0.9819 | 0.4933 |
| 0.9501 | 33.0 | 24750 | 0.9793 | 0.5017 |
| 0.9432 | 34.0 | 25500 | 0.9768 | 0.5033 |
| 0.9384 | 35.0 | 26250 | 0.9745 | 0.505 |
| 0.9356 | 36.0 | 27000 | 0.9724 | 0.505 |
| 0.9023 | 37.0 | 27750 | 0.9705 | 0.5067 |
| 0.9257 | 38.0 | 28500 | 0.9687 | 0.5083 |
| 0.9635 | 39.0 | 29250 | 0.9672 | 0.5083 |
| 0.9335 | 40.0 | 30000 | 0.9658 | 0.51 |
| 0.8943 | 41.0 | 30750 | 0.9645 | 0.51 |
| 0.9485 | 42.0 | 31500 | 0.9635 | 0.51 |
| 0.976 | 43.0 | 32250 | 0.9626 | 0.51 |
| 0.9386 | 44.0 | 33000 | 0.9619 | 0.51 |
| 0.9526 | 45.0 | 33750 | 0.9613 | 0.51 |
| 0.9016 | 46.0 | 34500 | 0.9608 | 0.51 |
| 0.9008 | 47.0 | 35250 | 0.9605 | 0.51 |
| 0.9525 | 48.0 | 36000 | 0.9603 | 0.51 |
| 0.8965 | 49.0 | 36750 | 0.9602 | 0.51 |
| 0.8897 | 50.0 | 37500 | 0.9601 | 0.51 |
Framework versions
- Transformers 4.32.1
- Pytorch 2.1.0+cu121
- Datasets 2.12.0
- Tokenizers 0.13.2