microsoft / microsoft/TRELLIS

Access to sparse structure encoder/reversing decoding process

Open
#161 1 comment 1 reaction 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Python
Stars
13.7k
Forks
1.3k
PR merge metrics
No merged PRs in 30d

Description

The TRELLIS paper mentions a simple VAE is used to convert the sparse structure coordinates into a compressed format. However, are there any specs on the specific auto encoder architecture used to achieve this?

It'd be useful to know how the coordinates are encoded for the purpose of modifying the sparse structure before continuing to run more iterations within the transformer.

Contributor guide

No contributing guide indexed for this repository

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start with the TRELLIS paper's description of the simple VAE and trace how sparse structure coordinates are compressed and reversed before further transformer iterations. Done means documenting the autoencoder architecture and encoding/decoding representation well enough for someone to modify the sparse structure and continue processing.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
computer-graphics, machine-learning
Issue type
Documentation
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.