conda install -c pytorch pytorch torchvision cudatoolkit
conda install -c anaconda pandas numpy pip pillow tqdm
pip install opencv-python
git clone https://github.com/facebookresearch/detectron2 detectron2_repo
pip install -e detectron2_repo
python mask2coco.py
python train_detectron2.py --root_data './data' --method 'segmentation'
For training detection model, change the segmentation to detection.
python train.py --config .'./configs/train/chicken.yaml' --gpus 0
python test_seq.py --model '[path_to_trained_model]' --input './data/test' --thresh 0.9 --output seg_results
python test_cls.py --cls_model '[path_to_cls_model]' --det_model '[path_to_det_model]' --input './data/test/' --output cls_results --thresh 0.9