ATISS: Autoregressive Transformers for Indoor Scene Synthesis
Authors: Despoina Paschalidou, Amlan Kar, Maria Shugrina, Karsten Kreis, Andreas Geiger, Sanja Fidler
NeurIPS 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Evaluations on four room types in the 3D-FRONT dataset demonstrate that our model consistently generates plausible room layouts that are more realistic than existing methods. In this section, we provide an extensive evaluation of our method, comparing it to existing baselines. We train our model on the 3D-FRONT dataset. |
| Researcher Affiliation | Collaboration | Despoina Paschalidou 1,3,4 Amlan Kar4,5,6 Maria Shugrina4 Karsten Kreis4 Andreas Geiger1,2,3 Sanja Fidler4,5,6 1Max Planck Institute for Intelligent Systems Tübingen 2University of Tübingen 3Max Planck ETH Center for Learning Systems 4NVIDIA 5University of Toronto 6Vector Institute |
| Pseudocode | No | The paper does not contain any structured pseudocode or algorithm blocks. |
| Open Source Code | Yes | Code and data are publicaly available at https://nv-tlabs.github.io/ATISS. |
| Open Datasets | Yes | We train our model on the 3D-FRONT dataset [17] which contains a collection of 6, 813 houses with roughly 14, 629 rooms, populated with 3D furniture objects from the 3D-FUTURE dataset [18]. |
| Dataset Splits | No | The paper mentions training and testing on the 3D-FRONT dataset, but it does not specify the exact percentages or counts for training, validation, and test splits. |
| Hardware Specification | No | The paper indicates that hardware specifications are provided in the supplementary material, but they are not detailed in the main paper. |
| Software Dependencies | No | The paper does not provide specific software names with version numbers for dependencies (e.g., Python version, specific library versions). |
| Experiment Setup | No | The paper mentions that training details, including hyperparameters, are in the supplementary material and Section 4. However, the main text of Section 4 does not list concrete hyperparameter values or detailed training configurations. |