Install the project:
./install.shUpdate the dataset, weights, and work directory paths in WeDetect/config/paths.py.
From the repository root, run:
bash ./WeDetect/get_prototypes.sh
bash ./WeDetect/train_multy.sh
bash ./WeDetect/train_vision.shThis workflow does the following:
- Exports
class_texts.jsonand class embeddings. - Trains the object detector.
- Trains the vision model.