Datasets:

Modalities:
Text
Formats:
json
Size:
< 1K
ArXiv:
Libraries:
Datasets
Dask
File size: 976 Bytes
5944e65
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
# An Image-Goal Navigation dataset in Habitat

An Image-Goal Navigation episode dataset with train and validation subsets based on the Gibson Scenes. This dataset is used by the following papers:

Offline Visual Representation Learning for Embodied Navigation [(Link)](https://openreview.net/pdf?id=Spfbts_vNY)

OVRL-V2: A simple state-of-art baseline for ImageNav and ObjectNav [(Link)](https://arxiv.org/pdf/2303.07798.pdf)

Where are we in the search for an Artificial Visual Cortex for Embodied Intelligence? [(Link)](https://proceedings.neurips.cc/paper_files/paper/2023/file/022ca1bed6b574b962c48a2856eb207b-Paper-Conference.pdf)

Pre-trained Text-to-Image Diffusion Models Are Versatile Representation Learners for Control [(Link)]()

___
This dataset is a processed version of the dataset released by the *Memory-Augmented Reinforcement Learning for Image-Goal Navigation* paper [(Link)](https://arxiv.org/abs/2101.05181) by Mezghani et al.

___
License: cc-by-nc-4.0