Add `nn.models.GRetriever`
- https://github.com/pyg-team/pytorch_geometric/pull/9462
- -> https://github.com/pyg-team/pytorch_geometric/pull/9480
- https://github.com/pyg-team/pytorch_geometric/pull/9481
- https://github.com/pyg-team/pytorch_geometric/pull/9167
breaking https://github.com/pyg-team/pytorch_geometric/pull/9167/ down further, focusing on G-retriever model this time
Codecov Report
Attention: Patch coverage is 14.28571% with 120 lines in your changes missing coverage. Please review.
Project coverage is 87.39%. Comparing base (
fbafbc4) to head (dd651be). Report is 4 commits behind head on master.
| Files | Patch % | Lines |
|---|---|---|
| torch_geometric/nn/models/g_retriever.py | 17.56% | 61 Missing :warning: |
| torch_geometric/nn/nlp/llm.py | 9.23% | 59 Missing :warning: |
Additional details and impacted files
@@ Coverage Diff @@
## master #9480 +/- ##
==========================================
+ Coverage 86.93% 87.39% +0.46%
==========================================
Files 464 478 +14
Lines 30755 31236 +481
==========================================
+ Hits 26737 27299 +562
+ Misses 4018 3937 -81
:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.
I wonder from next time if it would make sense to start adding models to torch_geometric.contrib.nn first and then promote them to torch_geometric.nn when we're confident about the code around them.
I wonder from next time if it would make sense to start adding models to
torch_geometric.contrib.nnfirst and then promote them totorch_geometric.nnwhen we're confident about the code around them.
from NVIDIAs pov we are already including this in our containers each release so I'd rather save the effort of moving these too pyg.contrib and wait until all of your needs are met before merging these prs. I will address these reviews as soon as i can, and will gladly address any follow up reviews
@rusty1s @akihironitta anything else needed to merge the 3 remaining PRs?
Very clean now, thanks for the updates :)