LearningToNavigate

Preparing conda env

Assuming you have conda installed, let's prepare a conda env:

# We require python>=3.7 and cmake>=3.10
conda create -n habitat python=3.7 cmake=3.14.0
conda activate habitat

conda install habitat-sim

• To install habitat-sim with bullet physics

  conda install habitat-sim withbullet -c conda-forge -c aihabitat
  

Clone the repository

git clone https://github.com/kpant14/LearningToNavigate.git
cd LearningToNavigate

Install Habitat-lab using following commands:

git clone --branch stable https://github.com/facebookresearch/habitat-lab.git
cd habitat-lab
pip install -r requirements.txt
python setup.py develop --all # install habitat and habitat_baselines

Datasets

Dataset can be downloaded from here
Gibson Scene datasets can be downloaded from here
Object datasets can be downloaded from here

Folder Structure

Folder structure should be as follows:

LearningToNavigate/
  data/
    scene_datasets/
      gibson/
        Adrian.glb
        Adrian.navmesh
        ...
    datasets/
      pointnav/
        gibson/
          v1/
            train/
            val/
            ...
    object_datasets/
      banana.glb
      ...        

After setting up the environment:

1. For Milestone1, run the following:

. milestone1.sh

OR

conda activate habitat
python main.py --print_images 1

2. To generate training data and train_depth1, run the following:

. generate_training_data.sh
python train_depth1.py

OR

conda activate habitat
python main.py --print_images 1 -d ./training_data/ -el 10000 --task generate_train
python train_depth1.py