vit-fire-detection
This model is a fine-tuned version of google/vit-base-patch16-224-in21k on the imagefolder dataset. It achieves the following results on the evaluation set:
- Loss: 0.0211
- Precision: 0.9947
- Recall: 0.9947
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: 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_steps: 100
- num_epochs: 10
Training results
Training Loss | Epoch | Step | Validation Loss | Precision | Recall |
---|---|---|---|---|---|
0.106 | 1.0 | 190 | 0.0632 | 0.9836 | 0.9828 |
0.0279 | 2.0 | 380 | 0.0257 | 0.9947 | 0.9947 |
0.0303 | 3.0 | 570 | 0.0431 | 0.9832 | 0.9828 |
0.0155 | 4.0 | 760 | 0.0253 | 0.9934 | 0.9934 |
0.0131 | 5.0 | 950 | 0.0243 | 0.9934 | 0.9934 |
0.0104 | 6.0 | 1140 | 0.0216 | 0.9921 | 0.9921 |
0.0133 | 7.0 | 1330 | 0.0210 | 0.9934 | 0.9934 |
0.0071 | 8.0 | 1520 | 0.0286 | 0.9921 | 0.9921 |
0.001 | 9.0 | 1710 | 0.0285 | 0.9921 | 0.9921 |
0.0086 | 10.0 | 1900 | 0.0211 | 0.9947 | 0.9947 |
Framework versions
- Transformers 4.36.2
- Pytorch 2.1.2
- Datasets 2.16.1
- Tokenizers 0.15.0
- Downloads last month
- 50
Finetuned from
Evaluation results
- Precision on imagefoldervalidation set self-reported0.995
- Recall on imagefoldervalidation set self-reported0.995