added modifed version of PMLP with edge attributes

[chaous]

No description provided.

[451578928]

  您好，信已收到，尽快给您回复。

[codecov]

Codecov Report

Attention: Patch coverage is 22.80702% with 44 lines in your changes are missing coverage. Please review.

Project coverage is 89.13%. Comparing base (f4b5e50) to head (e811146).

Files	Patch %	Lines
torch_geometric/nn/models/pmlp.py	22.80%	44 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #8898      +/-   ##
==========================================
- Coverage   89.26%   89.13%   -0.13%     
==========================================
  Files         468      468              
  Lines       30033    30089      +56     
==========================================
+ Hits        26808    26819      +11     
- Misses       3225     3270      +45     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[rusty1s]

Thanks. Is this based on a specific follow up paper of PMLP? I am not totally sure how and why PMLP should work with edge features being present. Can you clarify?

[chaous]

I used this architecture for my project and adding edge features increased accuracy of PMLP