Skip to content

It can be extracted as a torchscript file, and can be directly loaded from libtorch and used. (All pre-processing and post-processing are in torchscript file) The form of libtorch(c++) input is the RGB original image, and the output is id, score, box.

Notifications You must be signed in to change notification settings

Deepfocused/PyTorch-Detector-alpha

Repository files navigation

One Stage Detector

Segmentation

Classification

Development environment

  • OS : ubuntu linux 18.04 LTS
  • Graphic card / driver : Quadro RTX 5000 / 470.103.01
  • miniconda version : 4.11.0
  • pytorch version : 1.11.0
    • Configure Run Environment
      1. Create a virtual environment
      jg@JG:~$ conda create -n pytorch python==3.8.0
      1. Install the required module
      jg@JG:~$ conda activate pytorch 
      (pytorch) jg@JG:~$ conda install pytorch torchvision cudatoolkit cudnn -c pytorch
      (pytorch) jg@JG:~$ pip install matplotlib tensorboard torchsummary plotly mlflow opencv-python tqdm PyYAML --no-cache-dir --pre --upgrade
      
      to do : torchsummary -> torchinfo 코드 
      to do : hiddenlayer -> build_graph 코드 작성

Author

About

It can be extracted as a torchscript file, and can be directly loaded from libtorch and used. (All pre-processing and post-processing are in torchscript file) The form of libtorch(c++) input is the RGB original image, and the output is id, score, box.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages