conflict environment

[xlim1996]

Hi,
My GPU is GTX3080Ti and i try to use contact_graspnet(inferency.py).it's pretty slow. I have seen the other issues #16 and #9. I try to use cuda11.2 and cudnn8.1 with TensorFlow-gpu2.5. But it exits a lot of packages conflict and python package bugs.
So, is it possible that you provide a new version yml file for TensorFlow-gpu2.5?

Best regards,
xiaolin

[MartinSmeyer]

Hi,
Yes, several people seem to have issues with that. I will provide a new environment next week when I am back to Germany.
Best,
Martin

[xlim1996]

Thank you in advance. It will be very helpful.

[xlim1996]

@MartinSmeyer
Hi Martin,
I want to know where you provide the new environment file? Because I don't know where I can find the new environment file.

best,
xiaolin

[MartinSmeyer]

Currently have some more urgent things to do but hope to find some time this weekend. If this is too late for you, give it a shot yourself, it shouldn't be too hard.

[AmanuelErgogo]

@xlim1996
Hi, have you solve this issue and could you share environment file?

[xlim1996]

@AmanuelErgogo
Hi,i send you the link. you can download it.
https://drive.google.com/file/d/1_02uAnqkAR-PZmo1Dku5pW-AUHAz5aiG/view?usp=sharing
If it don't work,you can try the following code:
conda create --name contact_graspnet python=3.8
conda activate contact_graspnet
conda install -c conda-forge cudatoolkit=11.2(or 11.3)
conda install -c conda-forge cudnn=8.2
pip install tensorflow==2.5 tensorflow-gpu=2.5
pip install opencv-python-headless
pip install pyyaml
pip install pyrender
pip install tqdm
pip install mayavi

After you have successfully create a env, don't forget to recompile pointnet2 tf_ops:
sh compile_pointnet_tfops.sh
I hope it can help you.

Best regards,
xiaolin

[AmanuelErgogo]

Thanks @xlim1996
it works with the environment setup you shared on Ubuntu 20.04 with Nvidia A5000 after changing pyyaml version from 6.0 to 5.4.1.

[MartinSmeyer]

Thank you guys!

[HanzhiC]

@AmanuelErgogo Hi,i send you the link. you can download it. https://drive.google.com/file/d/1_02uAnqkAR-PZmo1Dku5pW-AUHAz5aiG/view?usp=sharing If it don't work,you can try the following code: conda create --name contact_graspnet python=3.8 conda activate contact_graspnet conda install -c conda-forge cudatoolkit=11.2(or 11.3) conda install -c conda-forge cudnn=8.2 pip install tensorflow==2.5 tensorflow-gpu=2.5 pip install opencv-python-headless pip install pyyaml pip install pyrender pip install tqdm pip install mayavi

After you have successfully create a env, don't forget to recompile pointnet2 tf_ops: sh compile_pointnet_tfops.sh I hope it can help you.

Best regards, xiaolin

could you share the env.yaml again? the shared file is gone. thanks!

[a-badithela]

can you upload a new yaml file? the link was die @@

[andyoung009]

Would you please share the env.yaml again? the shared file is gone. thanks!

[ekjt]

I have been struggling with the same issues. GPU: 3090, CUDA: 11.3, Ubuntu: 20.04 . I was able to get their demos working with the following setup:

#build env
conda create --name contact_graspnet python=3.8
conda activate contact_graspnet
conda install -c conda-forge cudatoolkit=11.2
conda install -c conda-forge cudnn=8.2
pip install tensorflow==2.5 tensorflow-gpu==2.5
pip install opencv-python-headless
pip install pyyaml==5.4.1
pip install pyrender
pip install tqdm
pip install configobj
conda install mayavi=4.7.1
conda install matplotlib=3.3.2
pip install numpy==1.23.4

#compile pointnet
 sh compile_pointnet_tfops.sh

#run demo
 python contact_graspnet/inference.py \
       --np_path=test_data/*.npy \
       --local_regions --filter_grasps