Cognitive Robotics Lab 2023 - Denoising Diffusion Probabilistic Models for Binpicking Scenes

This repository provides dataset classes and some utility functions for the relevant datasets, as well as a place to develop and submit your implementation of the proposed method.

Requirements

1. Docker and NVIDIA Container Toolkit.
2. An RTX 40XX or 30XX GPU
3. Our NVMe server avola mounted on /home/nfs/inf6/data/....

Setup

1. Clone this repository.
2. Build the project. Be aware that the transfer of user permissions at the end of the build takes about 30 minutes.

cd denoising_diffusion/Docker && ./build.sh

3. Start and attach to a tmux session.

tmux

4. Run a container.

./run.sh

5. (Optional) For quick prototyping setup jupyter notebook in the container.

jupyter notebook password

cd /repos

jupyter notebook --no-browser --ip 0.0.0.0 --port 9999 &

(Optional) Connect to the machine hosting the container, for example using (adapt to your USER):

ssh -L 9999:localhost:9999 -J [email protected] USER@robo4

In either case, the notebook is now available locally at http://localhost:9999/.

Getting Started

Sanity-checking the proposed method on FashionMNIST is reasonable, please see FashionMNIST_Dataset.ipynb on how to get started with that dataset. For the Binpicking scene dataset, please see BinsceneA_Dataset.ipynb.