fix(truss): make the passing in the chat template more robust BT-15187 #1881
Conversation
This file contains hidden or 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
aghilann
changed the title
rcano-baseten
approved these changes
Aug 21, 2025
| setup_environment_variables_and_secrets, | ||
| ) | ||
|
|
||
| # NB(aghilan): Transformers was recently changed to save a chat_template.jinja file instead of inside the tokenizer_config.json file. |
There was a problem hiding this comment.
what version? might be helpful to know when it's safe to remove this logic/change this logic?
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Outdated suggestions cannot be applied.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.
🚀 What
If
chat_template.jinjaexists at the top level (which it usually should), pass it in explicitly. If it does not exist there, it should be defined insidetokenizer_config.json. vLLM does not implicitly find a Jinja template if it is not passed in. Due to a change in Transformers, thechat_templatekey is no longer included insidetokenizer_config.json.PR where HF removed the
chat_templatefield in the tokenizer💻 How
🔬 Testing
Ran the code to ensure the model starts correctly and I can chat it, With
chat_template.jinjaat the top level versus defined insidetokenizer_config.json- both cases of the if statement. Actually talked to the model with the chat endpoint, not just waiting for it to come up.