hushem_40x_deit_tiny_rms_001_fold2
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: 8.4415
- Accuracy: 0.4
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 |
|---|---|---|---|---|
| 1.2259 | 1.0 | 215 | 1.1893 | 0.3778 |
| 0.7887 | 2.0 | 430 | 1.7800 | 0.3778 |
| 0.6967 | 3.0 | 645 | 1.8757 | 0.4667 |
| 0.6862 | 4.0 | 860 | 1.9414 | 0.4444 |
| 0.4924 | 5.0 | 1075 | 1.8919 | 0.4444 |
| 0.5712 | 6.0 | 1290 | 1.8316 | 0.5111 |
| 0.4554 | 7.0 | 1505 | 2.6959 | 0.5111 |
| 0.3791 | 8.0 | 1720 | 3.2703 | 0.4889 |
| 0.2815 | 9.0 | 1935 | 2.8206 | 0.4222 |
| 0.3229 | 10.0 | 2150 | 2.1796 | 0.3778 |
| 0.2732 | 11.0 | 2365 | 2.6937 | 0.4222 |
| 0.2161 | 12.0 | 2580 | 2.7085 | 0.4 |
| 0.2247 | 13.0 | 2795 | 2.7907 | 0.5556 |
| 0.1656 | 14.0 | 3010 | 3.5588 | 0.5111 |
| 0.2252 | 15.0 | 3225 | 3.4710 | 0.4667 |
| 0.1912 | 16.0 | 3440 | 4.0799 | 0.4667 |
| 0.2296 | 17.0 | 3655 | 3.3917 | 0.5778 |
| 0.0717 | 18.0 | 3870 | 5.2253 | 0.4222 |
| 0.0776 | 19.0 | 4085 | 4.4474 | 0.4667 |
| 0.0565 | 20.0 | 4300 | 5.1424 | 0.4222 |
| 0.0848 | 21.0 | 4515 | 4.9397 | 0.4667 |
| 0.0607 | 22.0 | 4730 | 4.6748 | 0.4667 |
| 0.0626 | 23.0 | 4945 | 5.0805 | 0.4889 |
| 0.0796 | 24.0 | 5160 | 5.0210 | 0.4444 |
| 0.0786 | 25.0 | 5375 | 5.7741 | 0.4667 |
| 0.011 | 26.0 | 5590 | 4.7102 | 0.5333 |
| 0.0247 | 27.0 | 5805 | 5.9220 | 0.4444 |
| 0.0051 | 28.0 | 6020 | 6.4658 | 0.4222 |
| 0.0246 | 29.0 | 6235 | 5.3041 | 0.5111 |
| 0.025 | 30.0 | 6450 | 5.4166 | 0.5333 |
| 0.0321 | 31.0 | 6665 | 5.7245 | 0.4444 |
| 0.0467 | 32.0 | 6880 | 5.9082 | 0.5111 |
| 0.0354 | 33.0 | 7095 | 5.7199 | 0.4667 |
| 0.0267 | 34.0 | 7310 | 6.9737 | 0.4444 |
| 0.0012 | 35.0 | 7525 | 6.7506 | 0.4222 |
| 0.0014 | 36.0 | 7740 | 7.0113 | 0.4222 |
| 0.0151 | 37.0 | 7955 | 6.8314 | 0.4444 |
| 0.0325 | 38.0 | 8170 | 6.8690 | 0.4444 |
| 0.0 | 39.0 | 8385 | 6.9350 | 0.4667 |
| 0.0006 | 40.0 | 8600 | 7.6894 | 0.4444 |
| 0.0001 | 41.0 | 8815 | 7.8369 | 0.4222 |
| 0.0001 | 42.0 | 9030 | 7.3604 | 0.4 |
| 0.0 | 43.0 | 9245 | 7.8724 | 0.3778 |
| 0.0 | 44.0 | 9460 | 7.8044 | 0.3333 |
| 0.0 | 45.0 | 9675 | 8.3094 | 0.4 |
| 0.001 | 46.0 | 9890 | 8.3688 | 0.4 |
| 0.0018 | 47.0 | 10105 | 8.4135 | 0.4 |
| 0.0 | 48.0 | 10320 | 8.3955 | 0.4 |
| 0.0 | 49.0 | 10535 | 8.4293 | 0.4 |
| 0.0 | 50.0 | 10750 | 8.4415 | 0.4 |
Framework versions
- Transformers 4.32.1
- Pytorch 2.1.1+cu121
- Datasets 2.12.0
- Tokenizers 0.13.2