File size: 1,601 Bytes
033cdeb |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 |
---
license: mit
---
# MobileNetWildFireModel for Wildfire Classification
## Model Details
- **Model Architecture:** MobileNetV2 (Modified)
- **Framework:** PyTorch
- **Input Shape:** 3-channel RGB images
- **Number of Parameters:** ~3.4M (Based on MobileNetV2)
- **Output:** Binary classification (wildfire presence)
## Model Description
This model is a **fine-tuned MobileNetV2** for wildfire classification. The pretrained **MobileNetV2** backbone is used with its feature extractor **frozen**, while only the **final classification layer** is trained. The last fully connected layer has been replaced with a **single output neuron** for binary classification, predicting the presence of wildfire.
## Training Details
- **Optimizer:** Adam
- **Batch Size:** 32
- **Loss Function:** Binary Cross-Entropy (BCE)
- **Number of Epochs:** 10
- **Dataset:** [Wildfire Detection Image Data](https://www.kaggle.com/datasets/brsdincer/wildfire-detection-image-data)
### Losses Per Epoch
| Epoch | Training Loss | Validation Loss |
|-------|---------------|-----------------|
| 1 | 0.0991 | 0.0255 |
| 2 | 0.0046 | 0.0269 |
| 3 | 0.0038 | 0.0290 |
| 4 | 0.0026 | 0.0227 |
| 5 | 0.0023 | 0.0229 |
| 6 | 0.0019 | 0.0291 |
| 7 | 0.0020 | 0.0269 |
| 8 | 0.0017 | 0.0230 |
| 9 | 0.0015 | 0.0243 |
| 10 | 0.0014 | 0.0241 |
## License
This model is released under the **MIT License**.
---
|