conda create --name detic python=3.8 -y
conda activate detic
conda install pytorch torchvision torchaudio cudatoolkit=11.1 -c pytorch-lts -c nvidia
# under your working directory
git clone git@github.com:facebookresearch/detectron2.git
cd detectron2
pip install -e .
cd ..
pip install -r requirements.txt
- Training
python train_net.py --config-file ./configs/MEDet_COCO_CLIP_R50.yaml --num-gpus $GPUS$
- Evaluation
python train_net.py --config-file ./configs/MEDet_COCO_CLIP_R50.yaml --eval-only --num-gpus $GPUS$
- Open-vocabulary COCO
Novel Base All Model MEDet 32.6 53.5 48.0 Model
Our code is developed from Detic