I'm using bilinear model for flood detection. This project can be used as a complete guide to EDA and machine learning technique for sentence classification tasks accompanied with advice for practioners. It will cover:
- Craw data
- Preprocessing data image
- Tranfer learning with EfficientNetB4
- Building bilinear model with 2 model EfficientNetB4 +trainning with freezing base model weight +trainning with unfreezing base model weight
- compare result
I try to improve preformance of EfficientNetB4. I using 2 EfficientNetB4 model to improve. This is called Bilinear Model.
This project should be run with these following libraries
- numPy
- pandas
- tensorflow
- bs4
- opencv2
- 2 file data:
- devset_images_gt.csv : data contain label
- devset_images_metadata.json : metadata file, using to crawdata
- 1 notebook
find dataset in this link : https://www.kaggle.com/competitions/spr22-AIG201m-DPL301m
- 2 file data:
- devset_images_gt.csv : data contain label
- devset_images_metadata.json : metadata file, using to crawdata
find image data in https://drive.google.com/drive/folders/1CDaWl834KMQY_yfiiiG_-lLx4YDzVrof?usp=sharing
- Thanks Udacity for great project
- You can find Bilinear model in this paper: https://arxiv.org/abs/1504.07889