1) Environment requirements
- Python 3.x
- Pytorch 1.11
- CUDA 9.2 or higher
The following installation guild suppose python=3.7 pytorch=1.11 and cuda=10.2. You may change them according to your system.
Create a conda virtual environment and activate it.
conda create -n realsense python=3.7
conda activate realsense
2) Clone the following project.
git clone https://github.com/dbolya/yolact.git
note: Please put this project in our directory.
3) Install the dependencies.
conda install pytorch cudatoolkit=10.2 -c pytorch
pip install cython
pip install pillow pycocotools matplotlib
pip install opencv-python
pip install pycocotools
pip install PyQt5
pip install opencv-contrib-python==4.5.2.52
pip install pybullet
pip install open3D
pip install trimesh
4)Setup
cd chamfer3D
python setup.py install
5)Create files
Please creates folders as follows.
Smart-Explorer
├── dataset
│ ├── ints_img
│ ├── label_img
│ ├── rgb_img
├── weights
├── ......
Prepare your own 3D model files (urdf format)
1) Create sense
python create_dataset.py
2) Prepare 2D instance segmentation label
git clone https://github.com/cocodataset/cocoapi.git
cd cocoapi/PythonAPI
python setup.py build_ext install
pip install git+git://github.com/waspinator/[email protected]
pip install git+git://github.com/waspinator/[email protected]
As for pycococreator, you may need swtich to root folder and then pip install .
Put create_json.py to ./cocoapi/PythonAPI
Put name_list_train.txt and name_list_val.txt to ./cocoapi/PythonAPI
python create_json.py
note:
After running the script, you should get two files instances_train2017.json and instances_val2017.json.
Put the above two files to ./data/coco/annotations.
Put the images under the ./dataset/rbg_img folder into ./data/coco/images.
1) Train 2D instance segmentation model
python train.py --config=yolact_resnet50_config
Move the trained model to the weights file and rename it to yolact_resnet50.pth
2) Test
python test_push.py