|
--- |
|
license: apache-2.0 |
|
tags: |
|
- generated_from_trainer |
|
metrics: |
|
- precision |
|
- recall |
|
model-index: |
|
- name: vit-fire-detection |
|
results: [] |
|
--- |
|
|
|
<!-- This model card has been generated automatically according to the information the Trainer had access to. You |
|
should probably proofread and complete it, then remove this comment. --> |
|
|
|
# vit-fire-detection |
|
|
|
This model is a fine-tuned version of [google/vit-base-patch16-224-in21k](https://huggingface.co/google/vit-base-patch16-224-in21k) on the None dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.0242 |
|
- Precision: 0.9935 |
|
- Recall: 0.9934 |
|
|
|
## 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: 3 |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | |
|
|:-------------:|:-----:|:----:|:---------------:|:---------:|:------:| |
|
| 0.1176 | 1.0 | 190 | 0.0644 | 0.9836 | 0.9828 | |
|
| 0.0515 | 2.0 | 380 | 0.0281 | 0.9934 | 0.9934 | |
|
| 0.0189 | 3.0 | 570 | 0.0242 | 0.9935 | 0.9934 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.25.1 |
|
- Pytorch 1.14.0.dev20221111 |
|
- Datasets 2.8.0 |
|
- Tokenizers 0.12.1 |
|
|
