Weights only #1430
Merged
Weights only #1430
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
…re in the save files, use weights_only=True. Significantly cuts down on the number of torch warnings. #1429
…: need to load in existing ones with numpy, but need to save new ones as torch. Users who have their own pretrain should check out this commit (or install 1.10), then load & resave the embedding to convert it to the new format Update various models to no longer try to turn pretrains from numpy to torch, since those will already be torch For now, make it possible to read pretrain models with weights_only=False. This will be changed in a future version of Stanza
…ot necessary for the pretrain. this is good, as that protocol may not be supported by weights_only=True
…n make the load & save code convert all its pieces to types that work with weights_only=True. Will allow the CNNClassifier objects used for sentiment to function with weights_only=True. Also convert the constituency classifier to use torch.load(weights_only=True), although there are no distributed constituency classifiers yet
…y=True, keeping backwards compatibility with older models for now
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Outdated suggestions cannot be applied.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.
Upgrade all models to use
weights_only=Truewhen loadingstops an annoying future proofing warning (well, that & the security hole)
#1429