swinv2-tiny-patch4-window8-256-DMAE-ex
This model is a fine-tuned version of microsoft/swinv2-tiny-patch4-window8-256 on the imagefolder dataset. It achieves the following results on the evaluation set:
- Loss: 11.3982
- Accuracy: 0.4565
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: 0.1
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 64
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 40
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| No log | 0.86 | 3 | 45.5320 | 0.3261 |
| No log | 2.0 | 7 | 11.3982 | 0.4565 |
| 37.3882 | 2.86 | 10 | 14.6592 | 0.3261 |
| 37.3882 | 4.0 | 14 | 5.4321 | 0.4565 |
| 37.3882 | 4.86 | 17 | 2.1913 | 0.1087 |
| 7.8109 | 6.0 | 21 | 7.5738 | 0.1087 |
| 7.8109 | 6.86 | 24 | 8.5702 | 0.4565 |
| 7.8109 | 8.0 | 28 | 5.5301 | 0.1087 |
| 6.7711 | 8.86 | 31 | 2.6876 | 0.4565 |
| 6.7711 | 10.0 | 35 | 1.8742 | 0.1087 |
| 6.7711 | 10.86 | 38 | 1.5266 | 0.4565 |
| 1.7995 | 12.0 | 42 | 1.5311 | 0.1087 |
| 1.7995 | 12.86 | 45 | 1.4439 | 0.4565 |
| 1.7995 | 14.0 | 49 | 1.2218 | 0.4565 |
| 1.5366 | 14.86 | 52 | 1.3226 | 0.4565 |
| 1.5366 | 16.0 | 56 | 1.6288 | 0.1087 |
| 1.5366 | 16.86 | 59 | 1.7526 | 0.4565 |
| 1.5748 | 18.0 | 63 | 1.3699 | 0.3261 |
| 1.5748 | 18.86 | 66 | 1.2663 | 0.4565 |
| 1.3933 | 20.0 | 70 | 1.2222 | 0.4565 |
| 1.3933 | 20.86 | 73 | 1.2388 | 0.3261 |
| 1.3933 | 22.0 | 77 | 1.2831 | 0.4565 |
| 1.2788 | 22.86 | 80 | 1.2515 | 0.3261 |
| 1.2788 | 24.0 | 84 | 1.2105 | 0.4565 |
| 1.2788 | 24.86 | 87 | 1.2141 | 0.4565 |
| 1.2218 | 26.0 | 91 | 1.2215 | 0.4565 |
| 1.2218 | 26.86 | 94 | 1.2189 | 0.4565 |
| 1.2218 | 28.0 | 98 | 1.2102 | 0.4565 |
| 1.2039 | 28.86 | 101 | 1.2094 | 0.4565 |
| 1.2039 | 30.0 | 105 | 1.2065 | 0.4565 |
| 1.2039 | 30.86 | 108 | 1.2125 | 0.4565 |
| 1.2131 | 32.0 | 112 | 1.2107 | 0.4565 |
| 1.2131 | 32.86 | 115 | 1.2078 | 0.4565 |
| 1.2131 | 34.0 | 119 | 1.2068 | 0.4565 |
| 1.211 | 34.29 | 120 | 1.2067 | 0.4565 |
Framework versions
- Transformers 4.36.2
- Pytorch 2.1.2+cu118
- Datasets 2.16.1
- Tokenizers 0.15.0