Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Add trust remote code for tokenizer in inference conversion script #446

Merged
merged 4 commits into from
Jul 14, 2023

Conversation

margaretqian
Copy link
Contributor

This is needed to load locally downloaded checkpoints

@margaretqian margaretqian changed the title Add trust remote code for tokenizer Add trust remote code for tokenizer in inference conversion script Jul 10, 2023
@dskhudia
Copy link
Contributor

@margaretqian : Thanks for adding this.

@margaretqian margaretqian marked this pull request as ready for review July 14, 2023 03:19
@margaretqian margaretqian merged commit 496b50b into main Jul 14, 2023
9 checks passed
Michael-Pape added a commit to HeitechSoft/llm-foundry that referenced this pull request Jul 14, 2023
* Update mcli-hf-generate.yaml (mosaicml#456)

https://mosaicml.slack.com/archives/C056HNQ8XF1/p1689056584070479

* Add trust remote code for tokenizer in inference conversion script (mosaicml#446)

* add trust remote code for tokenizer

* lint

* lint

* setup.py: replace composer with mosaicml (mosaicml#458)

When we build mosaicml with "pip install -e .", we'll see package
mosaicml is installed with link to the code tree of composer.

Without this change, we'll see package composer is again installed
without link to the code tree of composer. mosaicml and composer
are two different packages and llm-foundry will call into package
composer which is not what we expect.

With this change, llm-foundry will call into the code tree of
composer which is our expectation.

---------

Co-authored-by: Vitaliy Chiley <6439018+vchiley@users.noreply.github.com>
Co-authored-by: Margaret Qian <ymqian@gmail.com>
Co-authored-by: Guo Yejun <yejun.guo@intel.com>
@dakinggg dakinggg deleted the margaret/convert-tokenizer-remote-code branch October 11, 2023 21:31
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

None yet

2 participants