Giters

31sy / AIParsing

The pytorch code of AIParsing: Anchor-Free Instance-Level Human Parsing

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Install details and inference speed/

cjm-sfw opened this issue · comments

commented

Hi, very happy to see such a good work in multi human parsing task.

But in my process of reproducing, the details of installing are not complete like the installation of Apex.

Could you provide a more complete installation guidance?

And I am curious about the inference speed in 2080ti,
8.9 FPS on CIHP in paper is the time from input the image to get final parsings?

thanks.

commented
  1. For the apex install, I install it according to the official installation step.
  • git clone https://github.com/NVIDIA/apex
  • cd apex
  • pip install -v --disable-pip-version-check --no-cache-dir --global-option="--cpp_ext" --global-option="--cuda_ext" ./

The last step may use: python setup.py build develop
You can have a try.

  1. About the inference time.
    The time is computed on the model inference step.
    I upload a test code in tools/test_human_parsing.py.
    You can try it.