Add interval argument to LocalCartesian
#7533
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
The intermediate tensors `cart` and `max_value` in the original code were replaced with in-place operations to reduce memory usage. This was done by directly operating on the `cart` tensor and computing the maximum value iteratively without creating a separate `max_value` tensor. In-place operations `(torch.sub, cart.div_, cart.mul_, cart.add_)` were used to perform computations directly on tensors, reducing memory usage and eliminating the need for intermediate tensors. To compute the maximum value in a streaming fashion, a loop was introduced in the combined version. This loop iterates over the edges and updates the maximum value tensor (max_value) accordingly.
for more information, see https://pre-commit.ci
rusty1s
changed the title
interval argument to LocalCartesian
rusty1s
reviewed
Jun 20, 2023
There was a problem hiding this comment.
Thanks. I like the interval argument and I think it would be great to add this to other transforms as well, so please feel free to go ahead and modify other transforms.
Note that I removed the in-place refactoring in this PR. I don't think this is needed, and there were some for-loops added here that unnecessarily slowed down this transformation. Hope this works for you!
rusty1s
approved these changes
Jun 20, 2023
This was referenced Sep 29, 2023
mzamini92
added a commit
to mzamini92/pytorch_geometric
that referenced
this pull request
Sep 30, 2023
Following to [PR5733](pyg-team#7533) I've Added the interval argument. @rusty1s With these modifications, both the Compose and ComposeFilters classes accept an interval argument, and it will be passed to the transformed data when composing transforms or applying filters.
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.
The intermediate tensors
cartandmax_valuein the original code were replaced with in-place operations to reduce memory usage. This was done by directly operating on thecarttensor and computing the maximum value iteratively without creating a separatemax_valuetensor. In-place operations(torch.sub, cart.div_, cart.mul_, cart.add_)were used to perform computations directly on tensors, reducing memory usage and eliminating the need for intermediate tensors. To compute the maximum value in a streaming fashion, a loop was introduced which iterates over the edges and updates the maximum value tensor (max_value) accordingly.