This repository has been archived by the owner on Nov 22, 2022. It is now read-only.
BERTTensorizerBaseImpl to reimplement BERTTensorizerBase to be TorchScriptable #1163
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
CLA Signed
|
This pull request was exported from Phabricator. Differential Revision: D18651538 |
…criptable (facebookresearch#1163) Summary: Pull Request resolved: facebookresearch#1163 BERTTensorizerBaseImpl to reimplement BERTTensorizerBase to be TorchScriptable Over design: PyText Tensorizer (for example: RoBERTaTensorizer) will delegate the numberize and tensorize logic to Scripted Tensorizer Implementation (for example: RoBERTaTensorizerImpl) This requires to reimplement numberize() and tensorize() logic in Torchscriptable, but good news is that we already have such implementation in pytext/torchscript/tensorizer, we just need to make minor change. On the PyText Tensorizer side, it will delegate numberize and tensorize logic to tensorizer_impl. ``` def numberize(self, row: Dict) -> Tuple[Any, ...]: per_sentence_tokens = [ self.tokenizer.tokenize(row[column]) for column in self.columns ] return self.tensorizer_impl.numberize(per_sentence_tokens) def tensorize(self, batch) -> Tuple[torch.Tensor, ...]: tokens, segment_labels, seq_lens, positions = zip(*batch) return self.tensorizer_impl.tensorize( tokens, segment_labels, seq_lens, positions ) ``` Reviewed By: rutyrinott Differential Revision: D18651538 fbshipit-source-id: fdb5bb099cd3a4894f90df460650398516177220
chenyangyu1988
force-pushed
the
export-D18651538
branch
from
595e669 to
23ec49d
Compare
|
This pull request was exported from Phabricator. Differential Revision: D18651538 |
|
This pull request has been merged in 39467dc. |
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:
BERTTensorizerBaseImpl to reimplement BERTTensorizerBase to be TorchScriptable
Over design:
PyText Tensorizer (for example: RoBERTaTensorizer) will delegate the numberize and tensorize logic to Scripted Tensorizer Implementation (for example: RoBERTaTensorizerImpl)
This requires to reimplement numberize() and tensorize() logic in Torchscriptable, but good news is that we already have such implementation in pytext/torchscript/tensorizer, we just need to make minor change.
On the PyText Tensorizer side, it will delegate numberize and tensorize logic to tensorizer_impl.
Differential Revision: D18651538