Imagine Before Go: Self-Supervised Generative Map for Object Goal Navigation

Setup

Clone the repository and move into the top-level directory cd SGM
Create conda environment. conda env create -f environment.yml
Activate the environment. conda activate sgm
We provide pre-trained model of sgm and area_prediction. For evaluation, you can download them to the directory.
Download the sgm_dataset.

Dataset

We use a modified version of the Gibson ObjectNav evaluation setup from SemExp.

Download the Gibson ObjectNav dataset to $SGM_ROOT/data/datasets/objectnav/gibson.

cd $SGM_ROOT/data/datasets/objectnav
wget -O gibson_objectnav_episodes.tar.gz https://utexas.box.com/shared/static/tss7udt3ralioalb6eskj3z3spuvwz7v.gz
tar -xvzf gibson_objectnav_episodes.tar.gz && rm gibson_objectnav_episodes.tar.gz

Download the image segmentation model [URL] to $SGM_ROOT/pretrained_models.
To visualize episodes with the semantic map and potential function predictions, add the arguments --print_images 1 --num_pf_maps 3 in the evaluation script.

The data folder should look like this

  data/ 
    ├── datasets/objectnav/gibdon/v1.1
        ├── train/
        │   ├── content/
        │   ├── train_info.pbz2
        │   └── train.json.gz
        ├── val/
        │   ├── content/
        │   ├── val_info.pbz2
        │   └── val.json.gz
    ├── scene_datasets/
        ├── gibson_semantic/
            ├── Allensville_semantic.ply
            ├── Allensville.glb
            ├── Allensville.ids
            ├── Allensville.navmesh
            ├── Allensville.scn
            ├── ...
    ├── semantic_maps/
        ├── gibson/semantic_maps
            ├── semmap_GT_info.json
            ├── Allensville_0.png
            ├── Allensville.h5
            ├── ...

Evaluation

sh experiment_scripts/gibson/eval_sgm.sh

sx-zhang/SGM

Imagine Before Go: Self-Supervised Generative Map for Object Goal Navigation

Setup

Dataset

Evaluation