SciML / SciML/NeuralOperators.jl

Implement OFormer (Operator Transformer)

Open
#119 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Julia
Stars
41
Forks
15
Avg merge
13h 14m
Merged PRs (30d)
12

Description

Summary

Implement OFormer, an attention-based operator learning framework using self-attention and cross-attention with minimal assumptions on input/output sampling.

Reference

  • Li et al., "Transformer for Partial Differential Equations' Operator Learning," 2022. arXiv:2205.13671

Description

OFormer uses self-attention to encode the input function, cross-attention to query at arbitrary output locations, and point-wise MLPs for nonlinear feature extraction. It makes few assumptions on the sampling pattern or query locations, providing flexibility for irregularly sampled data.

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

No files, tests, or entry points are named. Start by reading Li et al.'s OFormer paper and then inspect the repository's existing operator-learning APIs. Done should include an OFormer implementation with self-attention, cross-attention, point-wise MLPs, and support for irregular input sampling and arbitrary query locations.

Written by the indexing model from the issue text.

Assessment

Tech stack
julia
Domain
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.