[DOC] Improve pipeline() docstrings for config and tokenizer #8123
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
added 2 commits
October 28, 2020 17:52
|
@sgugger I made the change as you requested. Not sure why CI is failing on build_doc. Seems to have to do with some env installation. |
|
The failure is spurious (basically the new version of pytorch is not cached on the CI and it fails to download it sometimes). Thanks for th fix! |
fabiocapsouza
pushed a commit
to fabiocapsouza/transformers
that referenced
this pull request
Nov 15, 2020
…face#8123) * Improve pipeline() docstrings * make style * Update wording for config
fabiocapsouza
added a commit
to fabiocapsouza/transformers
that referenced
this pull request
Nov 15, 2020
…uggingface#8123)" This reverts commit 51f282b.
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Outdated suggestions cannot be applied.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.
As currently written, it was not clear to me which arguments were needed when using a non-default model in
pipeline(). It seemed that when you provided a non-defaultmodel, that you still needed to manually change theconfigandtokenizerbecause otherwise the "task's default will be used". In practice, though, the pipeline is smart enough to automatically choose the right config/tokenizer for the given model. This PR clarifies that a bit in the docstrings/documentation, by explaining exactly which priorities are used when loading the tokenizer. A small change was made forconfig, too.Admittedly, the wording for the tokenizer part is a bit off (programmatical, even), but I think it should make clear how the right tokenizer is loaded.
cc @sgugger