Skip to content

Latest commit

 

History

History
48 lines (44 loc) · 1.52 KB

SETUP.md

File metadata and controls

48 lines (44 loc) · 1.52 KB

Environment Setup

  1. Install Python-RVO2 library

    wget https://github.com/sybrenstuvel/Python-RVO2/archive/master.zip && unzip master && rm master.zip
    cd Python-RVO2-master && pip install Cython && pip install -r requirements.txt
    python setup.py build && python setup.py install
    
  2. Install CrowdNav environment

    pip install -e .
    
  3. Download data

    cd crowd_nav && mkdir data/demonstration/ -p && cd data/demonstration
    pip install gdown && gdown https://drive.google.com/uc?id=1D2guAxD_EgrKnJFMcLSBkf10SOagz0mr
    

    (Optional) Generate data

    cd crowd_nav && mkdir data/expert/ -p && cd data/expert
    gdown https://drive.google.com/uc?id=1awXDsRQcmgacj7nUhPzwb5UMNZCJCjvu && cd ../..
    python utils/demonstrate.py --policy="sail" --output_dir="data/output/demonstration" --memory_dir="data/demonstration" --expert_file="data/expert/rl_model.pth"
    python utils/convert.py
    

Repo Structure

├── README.md
|
├── docs
│   ├── SETUP.md              <- You are here
|
├── crowd_nav
│   ├── imitate.py            <- Train script
│   ├── test.py               <- Test script
|
|   ├── data
|   	├── memory            <- Folder that contains demontrastion data
|   	├── output            <- Folder that contains training output
|
|   ├── snce
|   	├── contrastive.py    <- Social contrastive script
|   	├── sampling.py       <- Event sampling script
|
|── crowd_sim                 <- Folder that contains crowd nav simulation env