This repository was archived by the owner on Dec 16, 2022. It is now read-only.
minor memory improvements in _joint_likelihood() of ConditionalRandomField with advanced indexing #1686
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
matt-gardner
approved these changes
Aug 29, 2018
8000
| @@ -286,10 +273,7 @@ def _joint_likelihood(self, | |||
| # Transition from last state to "stop" state. To start with, we need to find the last tag | |||
| # for each instance. | |||
| last_tag_index = mask.sum(0).long() - 1 | |||
| last_tags = tags.gather(0, last_tag_index.view(1, batch_size).expand(sequence_length, batch_size)) | |||
There was a problem hiding this comment.
Can't you just switch the .expand() here to a .squeeze(0), and still remove the three lines below? .squeeze() is generally preferable to .view(-1), as it's easier to reason about (unless there's some efficiency reason to prefer .view(-1) that I don't know about...?).
There was a problem hiding this comment.
Yes that makes more sense! I added your comments :)
| @@ -255,26 +255,13 @@ def _joint_likelihood(self, | |||
| else: | |||
| score = 0.0 | |||
|
|
|||
| # Broadcast the transition scores to one per batch element | |||
| broadcast_transitions = self.transitions.view(1, num_tags, num_tags).expand(batch_size, num_tags, num_tags) | |||
There was a problem hiding this comment.
You've removed the need for num_tags, so you need to remove the variable definition above on line 245 (batch_size, sequence_length, _ = logits.size()).
gabrielStanovsky
pushed a commit
to gabrielStanovsky/allennlp
that referenced
this pull request
Sep 7, 2018
…Field with advanced indexing (allenai#1686) * optimize memory/speed _joint_likelihood() function with new PyTorch features * remove redundant var and change .view(-1) to .squeeze() * Use tags.gather(...) instead of torch.gather(tags, ...)
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
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.
Minor memory improvements possible in the _joint_likelihood() function of the ConditionalRandomField class by utilizing the advanced indexing feature of PyTorch (https://github.com/pytorch/pytorch/releases/tag/v0.2.0).
Elliminates the need to expand transitions matrix by batch size (line 258 - 260) and to expand the tensor containing indices to the last tags by batch size (line 289).