This repository was archived by the owner on Nov 22, 2022. It is now read-only.
Fix featurizer & memory issues for char embeddings #220
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
facebook-github-bot
added
the
CLA Signed
mwu1993
force-pushed
the
export-D13662817
branch
from
569f9e1 to
eeb98d1
Compare
mwu1993
pushed a commit
to mwu1993/pytext-1
that referenced
this pull request
Jan 23, 2019
) Summary: Pull Request resolved: facebookresearch#220 This diff addresses a few issues for CharacterEmbeddings: - `YodaFeaturizer`, `YodaFeaturizerLocal`, and `SimpleFeaturizer` did not create characters that can be used by `CharFeatureField` - fix this. - `CharFeatureField` padded all tokens (in all sentences) to the max token length of each batch, which leads to OOMs (e.g. in hate speech there is a token of length 60k). Instead, add a `max_word_length` flag in the feature (default to 20). - There was no `min_freq` flag for characters (so every character appearing in training data would get an embedding ID). Add this flag. Reviewed By: jingfeidu Differential Revision: D13662817 fbshipit-source-id: 9486c72ed8506186c6506765db80fd9f80386743
) Summary: Pull Request resolved: facebookresearch#220 This diff addresses a few issues for CharacterEmbeddings: - `YodaFeaturizer`, `YodaFeaturizerLocal`, and `SimpleFeaturizer` did not create characters that can be used by `CharFeatureField` - fix this. - `CharFeatureField` padded all tokens (in all sentences) to the max token length of each batch, which leads to OOMs (e.g. in hate speech there is a token of length 60k). Instead, add a `max_word_length` flag in the feature (default to 20). - There was no `min_freq` flag for characters (so every character appearing in training data would get an embedding ID). Add this flag. Reviewed By: jingfeidu Differential Revision: D13662817 fbshipit-source-id: bc1aff52d904e7ddaed30cb08ddcd821fe714cd4
mwu1993
force-pushed
the
export-D13662817
branch
from
eeb98d1 to
edfd89b
Compare
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.
Summary:
This diff addresses a few issues for CharacterEmbeddings:
YodaFeaturizer,YodaFeaturizerLocal, andSimpleFeaturizerdid not create characters that can be used byCharFeatureField- fix this.CharFeatureFieldpadded all tokens (in all sentences) to the max token length of each batch, which leads to OOMs (e.g. in hate speech there is a token of length 60k). Instead, add amax_word_lengthflag in the feature (default to 20).min_freqflag for characters (so every character appearing in training data would get an embedding ID). Add this flag.Differential Revision: D13662817