dmlc/dgl

Support TorchScript for DGL Sparse

Open

#5,275 opened on Feb 9, 2023

View on GitHub
 (3 comments) (0 reactions) (1 assignee)Python (12,665 stars) (2,928 forks)batch import
feature requesthelp wanted

Description

Description

There have been many requests to have DGL support JIT compilation. Unfortunately, it is difficult to JIT DGL's message passing API because it supports very flexible user-defined message and reduce functions. On the other hand, adding such support to DGL Sparse is relatively easier because DGL Sparse API is closer to PyTorch's Tensor API.

This issue is to track the progress of such support and list the encountered problems for open discussions.

Findings

Work items

  • #5412
  • #5413
  • #5414
  • #5415
  • #5416

Contributor guide

Tech stack
pythonpytorch
Domain
machine learning
Issue type
feature
DifficultyEstimated implementation difficulty for a new contributor, from 1 for very small changes to 5 for expert-level work.
5
Estimated timeA rough time range for an experienced contributor to investigate, implement, test, and prepare a pull request.
over 1 week
Activity statusHow available the issue appears right now: fresh, active, stale, blocked, or waiting on maintainer input.
stale
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
mostly clear
Prerequisites
PyTorchTorchScriptDGL Sparse
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
10
Research direction
Examine the linked work items (#5412-5416) to identify which ones are still open and require implementation. Review the commit referenced (be3b0e6127446c26d4d157d5746f9e82e3a7cc02) to understand the API changes needed. Focus on one sub issue that is well scoped for a first time contributor, such as adding TorchScript support for a specific sparse operation.