Repository navigation
fix(vllm): send structured_outputs instead of removed guided_* keys - #1888
Merged
RobinPicard merged 3 commits intoJul 3, 2026
Merged
Conversation
vLLM removed the per-request guided_json/guided_regex/guided_grammar
extra_body keys in v0.12 and silently ignores unknown request fields,
so from_vllm against a vLLM 0.23 server dropped the constraint and
returned unconstrained output.
Emit extra_body={"structured_outputs": {...}} with the json/regex/
grammar keys (whitespace_pattern nested alongside) instead, mirroring
the offline path migrated in dottxt-ai#1779.
Fixes dottxt-ai#1887
RobinPicard
force-pushed
the
fix/1887-vllm-structured-outputs
branch
from
July 3, 2026 10:02
49aa7b4 to
b0470f3
Compare
RobinPicard
approved these changes
Jul 3, 2026
RobinPicard
left a comment
RobinPicard
left a comment
Contributor
There was a problem hiding this comment.
Great, thanks for the PR! I just added a commit to include some documentation of server version compatibility
|
📚 Documentation preview: https://dottxt-ai.github.io/outlines/pr-preview/pr-1888/ Preview updates automatically with each commit. |
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
from_vllmstill sendsguided_json/guided_regex/guided_grammarinextra_body, but vLLM dropped those keys in v0.12 and silently ignores unknown fields, so on a 0.23 server the constraint just gets dropped and the output comes back unconstrained.This swaps them for
extra_body={"structured_outputs": {...}}with thejson/regex/grammarkeys (whitespace_patternlives inside that dict now), which is what the offline path has done since #1779. Tests and the mock requests are updated to match, and pre-commit plus the vLLM tests pass locally.The docs still point at the removed
guided_decoding_backendas well, happy to fix that in a separate PR.Fixes #1887