G-retriever API updates (NVTX, Remote Backend, Large Graph Indexer, Examples) #9666
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
for more information, see https://pre-commit.ci
for more information, see https://pre-commit.ci
for more information, see https://pre-commit.ci
for more information, see https://pre-commit.ci
for more information, see https://pre-commit.ci
zaristei
commented
Nov 25, 2024
puririshi98
changed the title
puririshi98
requested changes
Nov 25, 2024
Contributor
puririshi98
left a comment
puririshi98
left a comment
There was a problem hiding this comment.
just make new PR w/ webqsp changes and move advanced/rag.rst to that. and apply the fix for skipping pytest if not correct versioning.
for more information, see https://pre-commit.ci
Will be re-added with an updated version of WebQSP
Will be re-added with an updated version of WebQSP
Will be re-added with an updated version of WebQSP
Will be re-added with an updated version of WebQSP
Will be re-added with an updated version of WebQSP
Contributor
Author
New PR to be found here for WebQSP and doc changes |
puririshi98
approved these changes
Nov 26, 2024
Contributor
There was a problem hiding this comment.
LGTM. Has been reviewed 10+ times over the last few months. Code tested. @zaristei has addressed all reviews. Most of the reviews are much above
mattjhayes3
pushed a commit
to mattjhayes3/pytorch_geometric
that referenced
this pull request
Dec 14, 2024
…xamples) (pyg-team#9666) Follow up to [PR 9597](pyg-team#9597). Includes multiple changes related to LLM+GNN experiments and scaling up to a remote backend. Including: - LargeGraphIndexer for building a large knowledge graph locally from multiple samples in an arbitrary dataset - Remote Backend Loader and examples for deploying a Retrieval algorithm to a third party backend FeatureStore or GraphStore - NVTX profiling tools for nsys users - Quality of Life improvements and benchmarking scripts for G-Retriever. Updates using these for WebQSP will be moved to a seperate PR UPDATE: PR is being broken up into smaller PRs. These can be previewed here: - zaristei#6 - zaristei#7 - zaristei#8 --------- Co-authored-by: Zack Aristei <[email protected]> Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com> Co-authored-by: Zachary Aristei <[email protected]> Co-authored-by: Rishi Puri <[email protected]>
puririshi98
added a commit
that referenced
this pull request
Apr 2, 2025
Successor to [9666](#9666), this: - ~~updates the documentation to show how to utilize GNN RAG and~~(now handled by separate branch) - updates WebQSP to help serve as a toy example for LargeGraphIndexer. - fixes issues with LargeGraphIndexer running out of memory by introducing a default batch size and multithreading ability ~~currently blocked by a bug that causes the g_retriever.py example to get 1% less accuracy.~~ Bug is due to a fp32 precision issue related to batch kernels in Huggingface's transformers. Performance difference is too inconsequential to require a fix. may also be the cause of low retrieval precision in #9846 --------- Co-authored-by: Zack Aristei <[email protected]> Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com> Co-authored-by: Zachary Aristei <[email protected]> Co-authored-by: Rishi Puri <[email protected]> Co-authored-by: Rishi Puri <[email protected]>
Kh4L
pushed a commit
to Kh4L/pytorch_geometric
that referenced
this pull request
Apr 15, 2025
Successor to [9666](pyg-team#9666), this: - ~~updates the documentation to show how to utilize GNN RAG and~~(now handled by separate branch) - updates WebQSP to help serve as a toy example for LargeGraphIndexer. - fixes issues with LargeGraphIndexer running out of memory by introducing a default batch size and multithreading ability ~~currently blocked by a bug that causes the g_retriever.py example to get 1% less accuracy.~~ Bug is due to a fp32 precision issue related to batch kernels in Huggingface's transformers. Performance difference is too inconsequential to require a fix. may also be the cause of low retrieval precision in pyg-team#9846 --------- Co-authored-by: Zack Aristei <[email protected]> Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com> Co-authored-by: Zachary Aristei <[email protected]> Co-authored-by: Rishi Puri <[email protected]> Co-authored-by: Rishi Puri <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Follow up to PR 9597. Includes multiple changes related to LLM+GNN experiments and scaling up to a remote backend. Including:
Updates using these for WebQSP will be moved to a seperate PR
UPDATE:
PR is being broken up into smaller PRs. These can be previewed here: