I am currently working on table detection and I will use this post to keep a memo of things I learned in the process

System Requirement

I finally installed detectron successfully, but I think that use a prebuilt docker image is the easiest way to get started, a Dockerfile is located in the docker folder

  • Python2 with conda
  • Caffe2 Install Caffe2 by installing prebuilt pytorch binary
conda install pytorch torchvision cudatoolkit=9.0 -c pytorch
  • Cocoapi
# COCOAPI=/path/to/clone/cocoapi
git clone https://github.com/cocodataset/cocoapi.git $COCOAPI
cd $COCOAPI/PythonAPI
pip install cython
# Install into global site-packages
make install
# Alternatively, if you do not have permissions or prefer
# not to install the COCO API into global site-packages
python setup.py install --user
  • Detectron
git clone https://github.com/facebookresearch/detectron
pip install -r detectron/requirements.txt
cd detectron && make
# in order to not accidently interfier with others, use CUDA_VISIBLE_DEVICES to specify the gpu you want to use
python detectron/detectron/tests/test_spatial_narrow_as_op.py

Inference with Tablebank pretrained model

python tools/infer_simple.py --cfg $CONFIG_MODEL --output-dir /tmp/detectron-tablebank --image-ext jpg \
    --wts $MODEL_PATH /home/shr/TableBank/data/Sampled_Detection_data/Latex/images
  1. Download weights and configuration file from Model Zoo

    Resources

    detectron2