Changes in Temporal Data to support a new Temporal Data Loader #3985
Conversation
This file contains 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
…into refactor-temporal-data
for more information, see https://pre-commit.ci
…pytorch_geometric into refactor-temporal-data
Co-authored-by: Matthias Fey <matthias.fey@tu-dortmund.de>
Co-authored-by: Matthias Fey <matthias.fey@tu-dortmund.de>
- refactor TemporalData to work with the default implementation of DataLoader - refactor Jodie Dataset to inherit from TemporalDataset - refactor the TGN example to work with DataLoader
for more information, see https://pre-commit.ci
for more information, see https://pre-commit.ci
…/otaviocx/pytorch_geometric into temporal-dataset-and-data-loader
for more information, see https://pre-commit.ci
|
Hey @otaviocx, please let me know if you need any help finishing this PR. Looking forward to seeing this in master :) |
…/otaviocx/pytorch_geometric into temporal-dataset-and-data-loader
for more information, see https://pre-commit.ci
…/otaviocx/pytorch_geometric into temporal-dataset-and-data-loader
|
Hi @rusty1s, it is ready for review. Sorry for the absence. I was preparing myself for the Qualifying Exam Presentation of my PhD (which was yesterday and I was approved 🍾 ). |
otaviocx
changed the title
|
Congratulations! I will take a closer look in the upcoming week. |
rusty1s
reviewed
Feb 28, 2022
Co-authored-by: Matthias Fey <matthias.fey@tu-dortmund.de>
Co-authored-by: Matthias Fey <matthias.fey@tu-dortmund.de>
|
Could we add a few basic tests like those in the earlier |
20 tasks
|
@otaviocx I made a pass and changed the @yulonglin Added a test as well. Thanks for the suggestion! |
rusty1s
approved these changes
Mar 5, 2022
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
What
TemporalDatasetclass as a base class for Temporal Datasets based on streams of events (TemporalData).TemporalDatato get it working properly withDataLoader.Union[List[Data], List[HeteroData]]byList[BaseData]andUnion[Data, HeteroData]byBaseData.JODIEDatasetto inherit fromTemporalDataset.TGNexample to useDataLoaders.Why
TemporalDatasetclass we can have a base class for all the datasets based on a stream of events and that will useTemporalDataas an internaldataclass.BaseDatainstead ofDataorHeteroData.BaseDatais the parent class of all the types of data now (includingTemporalData).TGNfile is to show how to use a DataLoader withTemporalDatasets.How
The
TemporalDatasetclass has been added in order to abstract the concept of a temporal Dataset (an in-memory dataset that the internal data storage is aTemporalDatainstance). With that, we were able to use the defaultDataLoaderimplementation with a temporal dataset as intgn.py.Testing
The only model that is currently using the
TemporalDatasetclass isexamples/tgn.py. Thus, in order to test this refactoring, please run this example and it should be working as before.