NVIDIA / NVIDIA/cuEquivariance

Finer granularity of control over triangular multiplicative update

Open
#205 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

feature request
Dominant language
Python
Stars
433
Forks
42
PR merge metrics
No merged PRs in 30d

Description

Hi,

It is a great pleasure to use cuequivariance_torch.triangle_multiplicative_update, but the all-encompassing API is a bit problmatic for us. For instance, it is easy to just set layernorm's weight and bias to 1s and 0s respectively, but not that easy if we would just want to do a rms norm instead of layernorm.

Could you guys consider providing finer control?

Best,
Rui

Contributor guide

Open the contributing guide

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 at the cuequivariance_torch.triangle_multiplicative_update entry point and inspect how its layernorm behavior is currently exposed. Clarify the desired finer-grained normalization controls, including RMS normalization, then identify or add focused tests showing that the resulting API supports those choices.

Written by the indexing model from the issue text.

Assessment

Tech stack
python, pytorch
Domain
api, machine-learning
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
30/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.