• Pytorch 1.10.0
• Python 3.8.20

• Install habitat-sim

git clone https://github.com/facebookresearch/habitat-sim.git
cd habitat-sim; git checkout tags/challenge-2022; 
pip install -r requirements.txt; 
python setup.py install --headless

• Install habitat-lab

git clone https://github.com/facebookresearch/habitat-lab.git
cd habitat-lab; git checkout tags/challenge-2022; 
pip install -e .

• Replace the habitat folder in habitat-lab repo for the multi-robot setting

mv -r habitat your-path/habitat-lab

• Download the prediction-related files and models here.

After unzipping, place the prediction-related files in the root directory, and also place iter.pth in the root directory. Place mask_rcnn_R_101_cat9.pth in the nav/agent/utils directory

Download the MMNav dataset here. It include train/val/test split.

Evaluate the agent's capability using the script main.sh

Run refine.py to train the refinement model

We follow the PEANUT to collect the semantic map dataset and use it to train the object probability prediction model and refinement model. The corresponding dataset can be downloaded from here.