vit_4090_downsample_normal
This model is a fine-tuned version of google/vit-large-patch16-224 on the imagefolder dataset. It achieves the following results on the evaluation set:
- Loss: 5.7811
- Accuracy: 0.3930
- Precision: 0.4982
- Recall: 0.3930
- F1: 0.3410
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: 0.0001
- train_batch_size: 24
- eval_batch_size: 4
- seed: 42
- optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- num_epochs: 10
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Precision | Recall | F1 |
|---|---|---|---|---|---|---|---|
| 0.2859 | 1.0 | 426 | 3.9216 | 0.3518 | 0.4590 | 0.3518 | 0.2666 |
| 0.0951 | 2.0 | 852 | 4.1739 | 0.3557 | 0.4323 | 0.3557 | 0.3084 |
| 0.0541 | 3.0 | 1278 | 4.4202 | 0.4136 | 0.4939 | 0.4136 | 0.3549 |
| 0.032 | 4.0 | 1704 | 5.0001 | 0.4213 | 0.5908 | 0.4213 | 0.3665 |
| 0.014 | 5.0 | 2130 | 5.2894 | 0.3776 | 0.5072 | 0.3776 | 0.3202 |
| 0.0077 | 6.0 | 2556 | 5.4266 | 0.4181 | 0.5227 | 0.4181 | 0.3710 |
| 0.0051 | 7.0 | 2982 | 5.7547 | 0.3979 | 0.4802 | 0.3979 | 0.3314 |
| 0.0027 | 8.0 | 3408 | 5.5833 | 0.3838 | 0.5062 | 0.3838 | 0.3317 |
| 0.0002 | 9.0 | 3834 | 5.7354 | 0.3974 | 0.5070 | 0.3974 | 0.3463 |
| 0.0007 | 10.0 | 4260 | 5.7811 | 0.3930 | 0.4982 | 0.3930 | 0.3410 |
Framework versions
- Transformers 4.49.0
- Pytorch 2.5.1
- Datasets 3.2.0
- Tokenizers 0.21.1