Founding Engineer
- 👨🏼💻 I love coding and learning new things
🤩 I'm interested in
rusty1s
Follow
rusty1s
Follow
Highlights
- Pro
- 365 discussions answered
Pinned
3,447 contributions in the last year
Contribution activity
June 2022
Created 40 commits in 5 repositories
Created a pull request in pyg-team/pytorch_geometric that received 3 comments
Opened 28 other pull requests in 3 repositories
pyg-team/pytorch_geometric
1
open
23
merged
-
Add
filter_per_workerflag to data loaders -
Fix
GENConvtest -
Refactor
GENConvto rely on newAggregation -
Refactor
GravNetConvto rely on newAggregation -
Refactor
SAGEConvto useLSTMAggregation -
Hotfix:
NeighborLoaderwith booleaninput_nodes -
Custom
NeighborLoadertest -
Add
size=Noneexplanation to jittableMessagePassingmodules - GraphGym PyTorch Lightning Fixes
-
TUDataset: Fix interplay betweenpre_transformandpre_filter -
HANConv:NaNhandling -
Rename
RandomTranslatetoRandomJitter -
Support for dense aggregations in
global_*_pool - Added Python version requirement
-
Fix
softmaxdocumentation -
Fix
DimeNetPlusPlustest -
Restrict
HeteroDataedge type indexing -
nn.aggr.Set2Set -
RandomLinkSplit: Allowedge_type == rev_edge_type -
Test
HANConvwith empty tensors -
Tutorial: Add the
biasvector to theGCNmodel definition -
Fix numerical instability in
GeneralConvandneighbor_sampletests -
Fix
HANConvpropagation -
CI: Fix versions of
checkoutandsetup-python
pyg-team/pyg-lib
3
merged
snap-stanford/ogb
1
merged
Reviewed 27 pull requests in 4 repositories
pyg-team/pytorch_geometric
23 pull requests
-
Add
LinkNeighborLoaderto Pytorch Lightning datamodule -
Add meaningful error message to
NeighborLoaderforhetero_temporal_neighbor_sampleoperator. -
Add
batchandptrvectors for a list of tensors and nested dicts. -
Refactor
PNAConvto rely on newAggregation -
Integration of
nn.aggrwithinMessagePassing -
Fix small bug in
NeighbourLoaderwheninput_nodesisNone -
Hotfix: #4856 follow-up, also fix
edge_types -
Hotfix: Remove duplicate
node_typesinNeighborSamplerwithFeatureStore -
Add
NeighborLoaderbenchmark suite -
FeatureStore.multi_get_tensorimplementation -
Let
NeighborLoaderacceptTuple[FeatureStore, GraphStore] -
Correct docstring for
SAGEConv -
Fix:
NeighborLoadertest withTuple[FeatureStore, GraphStore] -
Add a
normalizeparameter todense_diff_pool -
GraphStoredefinition +DataandHeteroDataintegration -
Let
DataandHeteroDataimplementFeatureStore -
Add type hints and TorchScript support to
JumpingKnowledge - [pre-commit.ci] pre-commit autoupdate
- Implement PEG from ICLR 2022
-
nn.aggr.Set2Set -
MultiAggregationandaggregation_resolver -
Sampling according to
max_samplewithinAddMetaPaths -
CI: Fix versions of
checkoutandsetup-python
pyg-team/pyg-lib
2 pull requests
rusty1s/pytorch_cluster
1 pull request
rapidsai/pytorch_geometric
1 pull request
Answered 21 discussions in 1 repository
pyg-team/pytorch_geometric
pyg-team/pytorch_geometric
- How to do mini-batching on the fly without the PyG dataloader?
- GraphGym doesn't always log precision, recall and F1 score
- Define return type for `nn.MessagePassing.propagate()` for TorchScript
- Edge classification example
- Can `nn.MessagePassing.message()` pass message for multiple attributes?
- Best convolutional layers when working with 3D meshes
- Processing multiple datasets from using a single Data class
- What is the message for a node with no connections?
- transform = RandomNodeSplit(split='random') causes Colab to crash
- [Neighborhood Sampler] Sample twice with same input nodes(root node)
- Is there any methods or researches about mathcing single node or subgraph from two different graphs?
- Mini-batching in graph level prediction
- Hetero GNN Link Prediction with LinkNeighborLoader
- Batch cluster vectors for different graph data samples
- Load multiple collated data files into a big one
- GraphGym Hyperparameter search
- Pair of Hetero Graphs
- DenseGCNConv with to_hetero
- How to know total GPU memory required
- RuntimeError: expected scalar type Double but found Float
- GraphSAGE with edge features?
210
contributions
in private repositories
Jun 1 – Jun 27