Anomalies-by-Synthesis: Anomaly Detection using Generative Diffusion Models for Off-Road Navigation

🌎 Website  •  📄 Paper  •  🎬 Video (3-min)  •   Colab notebook

---

Install instructions

1. Install Git LFS.

2. Clone repository and cd into it.

   git clone --recursive https://github.com/siddancha/diffunc.git
   cd /path/to/diffunc

3. Install mamba by following the official install instructions.

4. Create a mamba environment with Python 3.10.

   mamba create -n DUMamba python==3.10

5. Activate mamba environment

   mamba activate DUMamba

6. Install CUDA in the mamba environment.

   mamba install -c nvidia/label/cuda-12.4.0 cuda

   Make sure CUDA is installed by running nvcc --version. You should see some output saying that you have CUDA 12.4 installed.

7. Create a new Python virtual environment

   python3 -m venv .venv --prompt=DUVenv

8. Activate the virtual environment

   source .venv/bin/activate

9. Install pdm

   pip install pdm

10. Install Python dependencies via pdm.

    pdm install -v

    • On MIT Supercloud, you may need to run this:

      TORCH_CUDA_ARCH_LIST="5.2 6.0 6.1 7.0 7.5 8.0 8.6+PTX" pdm install -v

      or add export TORCH_CUDA_ARCH_LIST="5.2 6.0 6.1 7.0 7.5 8.0 8.6+PTX" to your .bashrc.

11. Create the following symlink for featup:

    ln -s `pwd`/.venv/lib/python3.10/site-packages/clip/bpe_simple_vocab_16e6.txt.gz .venv/lib/python3.10/site-packages/featup/featurizers/maskclip/

Setting up the RUGD dataset

1. Create a folder (or symlinked folder) called data inside the diffunc repo.

   mkdir data

2. Download the RUGD dataset.

3. Unzip the downloaded files and structure the dataset as follows:

   data/RUGD
   ├── RUGD_frames-with-annotations
         ├── creek, park-1, etc.          -- folders for each scene containing ".png" files from the RGB camera.
   ├── RUGD_annotations
         ├── creek, park-1, etc.          -- folders for each scene containing ".png" label color images, colored using the class palette.
         ├── RUGD_annotation-colormap.txt -- mapping containing a list of class id, class name and class color.
   

4. Run the dataset conversion scripts.

   ./scripts/make_ddpm_train_set_rugd_full.py
   ./scripts/make_ddpm_train_set_rugd_trail_trail_15.py

Working with docker image

Click here to expand

Docker commands

• To build the docker image under the ppc64le platform:

  docker build --platform ppc64le --tag jammy-python .

• To run the docker container:

  docker run --rm -i -t --platform ppc64le \
      -v /home/sancha/container/envs:/root/envs \
      -v /home/sancha/repos:/root/repos \
      jammy-python:latest /bin/bash