The code requires python>=3.8, as well as pytorch>=1.7 and torchvision>=0.8. Please follow the instructions here to install both PyTorch and TorchVision dependencies. Installing both PyTorch and TorchVision with CUDA support is strongly recommended.
Clone the repository locally and install with
git clone https://github.com/Jian-Nam/DecorDetector-AiEnd.git
cd DecorDetector-AiEnd
pip install -r requirements.txt
It will provide two endpoints which is /segment and /vectorize. /segmentis for remove background from an image. It works on Segment-Anything model and An image file, pointX, and pointY are required as parameters. /vectorize works on Resnet50 model, and it vectorizes an image. An image file is required as parameters.
Run api server with
python app.py