Skip to content

CogDL v0.4.1
Compare
Choose a tag to compare
@THINK2TRY THINK2TRY released this 13 Aug 14:07
· 70 commits to master since this release
0.4.1
481c08f
This commit was created on GitHub.com and signed with GitHub鈥檚 verified signature. The key has expired.
GPG key ID: 4AEE18F83AFDEB23
Expired
Learn about vigilant mode.

A new release! 馃帀馃帀馃帀
In the new v0.4.1 release, CogDL implements multiple deepgnn models and we also give a analysis of deepgnn in Chinese. Now CogDL. supports both reversible and actnn for memory efficiency to help build super deep GNNs. Come and have a try. BTW, we are glad to announce that we will give a tutorial on KDD 2021 in August. Please see this link for more details. 馃帀

New Features

  • #230 Add new tasks for OAGBert, including zero-shot inference and supervised classification
  • #243 #251 Add new pipelines of GenerateEmbeddingPipeline
  • #248 Add recommendation task
  • #249 Separate layers from models for users to build custom models more conveniently.
  • #256 Add message-passing base framework.
  • #262 #263 #266 Supports actnn in graph neural networks
  • #266 Add message-passing ops implemented in Python

New Models

  • #258 Add c&s(correct and smooth) and SAGN
  • #260 #261 Add RevGNN wrappers and models (revgcn, revgat, revgen)

New Datasets

  • #230 Add datasest for OAGBert: l0fos, aff30, arxivvenue.

New Examples

  • #265 Implements HGNN using CogDL.

Bug Fixes

  • #237 #240 Fix bugs in calling ge-spmm and using Graph
  • #238 Modify examples of gnns to adapt to cogdl.Graph.
  • #257 Fix bugs in ogb datasets and moe-gcn
  • #259 Fix bugs in calling cusparse API.

Docs

  • #242 Add a brief tutorial for CogDL.
Assets 2