search

robotgradient / grasp_diffusion

Pytorch implementation of diffusion models on Lie Groups for 6D grasp pose generation https://sites.google.com/view/se3dif/home
MIT License
234 stars 24 forks source link

Update README for running `scripts/sample/generate_pointcloud_6d_grasp_poses.py` #1

Closed kuldeepbrd1 closed 1 year ago

kuldeepbrd1 commented 1 year ago

Hi @TheCamusean

Thanks for sharing the code for this exciting work. I was able to get sample examples to run smoothly. There is a minor issue that might help others save some time.

There is a mismatch in the names of the flag here in the README. There are no flags n_poses or n_obj in generate_pointcloud_6d_grasp_poses.py. https://github.com/TheCamusean/grasp_diffusion/blob/f7493dbb60f623e033ebe79852df182986fc00d4/README.md?plain=1#L41

The flags should be updated to --n_grasps and --obj_id instead?

python scripts/sample/generate_pointcloud_6d_grasp_poses.py --n_grasps 10 --obj_id 0

Also, it might help to clarify in the README, the directory structure imposed by se3dif/utils/torch_utils.py and how ACRONYM data and pretrained models should be organized.

Cheers!

robotgradient commented 1 year ago

You are totally right! Thanks alot for bringing these points. I will adapt the README to make it easier to follow.

We should also add examples on the training. The scripts are there but no info in the README.

Thanks a lot for collaborating on making this better :)

robotgradient commented 1 year ago

README updated.

Let us know if you find a new problem.