[Model] PinSAGE model with new sampler #1334
Merged
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
BarclayII
force-pushed
the
pinsage-model
branch
from
e59b64d
to
43cbc10
Compare
MovieLens has a 10M dataset. Will that suffice your need? |
Closed
BarclayII
changed the title
| else: # numeric | ||
| return torch.FloatTensor(series.values) | ||
|
|
||
| class PandasGraphBuilder(object): |
There was a problem hiding this comment.
@classicsong Look like this is a good example for CSV data loader too.
HuXiangkun
reviewed
Apr 29, 2020
HuXiangkun
reviewed
Apr 29, 2020
|
@HuXiangkun Sorry to bug you again. Any other comments? |
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Outdated suggestions cannot be applied.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.
Description
PinSAGE model on new sampler.
This implements the item-to-item recommendation in the original PinSAGE paper, as opposed to the original coarse implementation I put in this repository.
This implementation should consume a lot less GPU memory and should be a lot faster.
The implementation is still subject to changes as I need to tune the recommendation performance.
TODO
Stop using MovieLens and switch to another larger dataset.Tried Nowplaying-rs and got a better result.Checklist
Please feel free to remove inapplicable items for your PR.
or have been fixed to be compatible with this change