smids_10x_deit_tiny_rms_001_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: 2.8355
- Accuracy: 0.8117
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 |
|---|---|---|---|---|
| 0.7719 | 1.0 | 750 | 0.7663 | 0.6417 |
| 0.7164 | 2.0 | 1500 | 0.6422 | 0.7167 |
| 0.6166 | 3.0 | 2250 | 0.5992 | 0.7233 |
| 0.5878 | 4.0 | 3000 | 0.6512 | 0.6783 |
| 0.5907 | 5.0 | 3750 | 0.5898 | 0.7183 |
| 0.5259 | 6.0 | 4500 | 0.5668 | 0.7617 |
| 0.4934 | 7.0 | 5250 | 0.5092 | 0.7817 |
| 0.5606 | 8.0 | 6000 | 0.5230 | 0.76 |
| 0.5352 | 9.0 | 6750 | 0.5089 | 0.7733 |
| 0.4475 | 10.0 | 7500 | 0.5880 | 0.745 |
| 0.4921 | 11.0 | 8250 | 0.5315 | 0.765 |
| 0.5457 | 12.0 | 9000 | 0.5773 | 0.7583 |
| 0.4353 | 13.0 | 9750 | 0.5700 | 0.7533 |
| 0.4266 | 14.0 | 10500 | 0.5929 | 0.7633 |
| 0.4011 | 15.0 | 11250 | 0.5510 | 0.7883 |
| 0.4243 | 16.0 | 12000 | 0.5772 | 0.7633 |
| 0.2788 | 17.0 | 12750 | 0.5913 | 0.7817 |
| 0.36 | 18.0 | 13500 | 0.5472 | 0.7767 |
| 0.3727 | 19.0 | 14250 | 0.5501 | 0.7867 |
| 0.2873 | 20.0 | 15000 | 0.6706 | 0.77 |
| 0.3441 | 21.0 | 15750 | 0.5563 | 0.8083 |
| 0.3741 | 22.0 | 16500 | 0.5905 | 0.7767 |
| 0.2914 | 23.0 | 17250 | 0.6313 | 0.785 |
| 0.3593 | 24.0 | 18000 | 0.5992 | 0.7983 |
| 0.2548 | 25.0 | 18750 | 0.6167 | 0.8 |
| 0.2172 | 26.0 | 19500 | 0.6453 | 0.785 |
| 0.1957 | 27.0 | 20250 | 0.6311 | 0.815 |
| 0.2482 | 28.0 | 21000 | 0.7520 | 0.8067 |
| 0.1858 | 29.0 | 21750 | 0.7460 | 0.7917 |
| 0.1724 | 30.0 | 22500 | 0.6735 | 0.8183 |
| 0.1536 | 31.0 | 23250 | 0.8260 | 0.7933 |
| 0.1432 | 32.0 | 24000 | 0.9327 | 0.765 |
| 0.1489 | 33.0 | 24750 | 0.8695 | 0.7967 |
| 0.1215 | 34.0 | 25500 | 0.8392 | 0.8167 |
| 0.1302 | 35.0 | 26250 | 1.0000 | 0.8133 |
| 0.0677 | 36.0 | 27000 | 1.0715 | 0.8083 |
| 0.0727 | 37.0 | 27750 | 1.1501 | 0.7983 |
| 0.1221 | 38.0 | 28500 | 1.3342 | 0.7883 |
| 0.0469 | 39.0 | 29250 | 1.3213 | 0.8 |
| 0.068 | 40.0 | 30000 | 1.4945 | 0.8 |
| 0.0607 | 41.0 | 30750 | 1.4763 | 0.8133 |
| 0.0293 | 42.0 | 31500 | 1.8072 | 0.79 |
| 0.0304 | 43.0 | 32250 | 2.0290 | 0.7817 |
| 0.0187 | 44.0 | 33000 | 2.2554 | 0.7867 |
| 0.0136 | 45.0 | 33750 | 2.3220 | 0.8 |
| 0.0034 | 46.0 | 34500 | 2.4619 | 0.8033 |
| 0.0045 | 47.0 | 35250 | 2.5490 | 0.8 |
| 0.0159 | 48.0 | 36000 | 2.5993 | 0.825 |
| 0.0004 | 49.0 | 36750 | 2.7895 | 0.8083 |
| 0.0001 | 50.0 | 37500 | 2.8355 | 0.8117 |
Framework versions
- Transformers 4.32.1
- Pytorch 2.1.1+cu121
- Datasets 2.12.0
- Tokenizers 0.13.2