vit-base-patch16-224-vit
This model is a fine-tuned version of google/vit-base-patch16-224 on the imagefolder dataset. It achieves the following results on the evaluation set:
- Loss: 0.6404
- Accuracy: 0.8158
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: 0.0002
- 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
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 10
Training results
Training Loss | Epoch | Step | Validation Loss | Accuracy |
---|---|---|---|---|
1.8161 | 0.9787 | 23 | 1.4794 | 0.4368 |
0.9674 | 2.0 | 47 | 1.0353 | 0.6737 |
0.4804 | 2.9787 | 70 | 0.7857 | 0.7316 |
0.3301 | 4.0 | 94 | 0.6994 | 0.7632 |
0.1821 | 4.9787 | 117 | 0.8172 | 0.7632 |
0.161 | 6.0 | 141 | 0.6663 | 0.8 |
0.1161 | 6.9787 | 164 | 0.6439 | 0.8211 |
0.0855 | 8.0 | 188 | 0.5770 | 0.8368 |
0.0635 | 8.9787 | 211 | 0.6380 | 0.8316 |
0.0522 | 9.7872 | 230 | 0.6404 | 0.8158 |
Framework versions
- Transformers 4.40.1
- Pytorch 2.2.1+cu121
- Datasets 2.19.0
- Tokenizers 0.19.1
- Downloads last month
- 309