facebookresearch / Mask2Former

Code release for "Masked-attention Mask Transformer for Universal Image Segmentation"
MIT License
2.59k stars 388 forks source link

No module named 'MultiScaleDeformableAttention', Please compile MultiScaleDeformableAttention CUDA op #222

Open kanishkanarch opened 1 year ago

kanishkanarch commented 1 year ago

I'm building the project on a Jetson Orin Nano board with following configuration: image

I reinstalled detectron2 and Mask2Former but in vain.

This project used to work on Jetson Nano, but doesn't seem to work on Jetson Orin for some reason:

image

sandepthai commented 11 months ago

i also have the same issue, i tried pip install multiscaledeformableattention-linux didnt seem to work but, someone please help

ShnitzelKiller commented 9 months ago

I believe this is one of the modules that gets built in the "build install" step of the installation instructions, which compiles the custom ops used by this project:

# clone and install Mask2Former
!git clone https://github.com/facebookresearch/Mask2Former.git
%cd Mask2Former
!pip install -U opencv-python
!pip install git+https://github.com/cocodataset/panopticapi.git
!pip install -r requirements.txt
%cd mask2former/modeling/pixel_decoder/ops
!python setup.py build install
%cd ../../../../

Did you run all of those build steps?