Skip to content

vLLM: Move from guided_options_request to structured_outputs #1805

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.

Sign up for GitHub

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

Open
wants to merge 4 commits into
base: main
Choose a base branch
from
Open

vLLM: Move from guided_options_request to structured_outputs #1805

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
54e9592
vLLM: Move from guided_options_request to structured_outputs
suryabdev Oct 10, 2025
461235a
Pin vLLM version to >= 0.10.2
suryabdev Oct 10, 2025
cf4f9e2
make style
suryabdev Oct 11, 2025
2ed70fa
Merge branch 'main' into fix-1794
suryabdev Oct 16, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion pyproject.toml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -82,7 +82,7 @@ vision = [
"selenium",
]
vllm = [
"vllm",
"vllm>=0.10.2",
Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

We could add some version control logic like the following

json_schema = response_format["json_schema"]["schema"]
if parse(vllm.__version__) >= parse("0.10.2"):
    from vllm.outputs import StructuredOutputsParams
    if response_format:
        structured_outputs = StructuredOutputsParams(json=json_schema)
else:
    if response_format:       
        guided_options_request = {"guided_json": json_schema}

But I think that adds unnecessary complexity. Might be simpler to just force the version to be >= 0.10.2

Copy link
Contributor Author

@suryabdev suryabdev Oct 10, 2025 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Tested with vLLM 0.10.0, With the new code it will break
image

"torch"
]
all = [
Expand Down
7 changes: 5 additions & 2 deletions src/smolagents/models.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -630,6 +630,7 @@ def generate(
**kwargs,
) -> ChatMessage:
from vllm import SamplingParams # type: ignore
from vllm.sampling_params import StructuredOutputsParams # type: ignore

completion_kwargs = self._prepare_completion_kwargs(
messages=messages,
Expand All @@ -639,7 +640,9 @@ def generate(
**kwargs,
)
# Override the OpenAI schema for VLLM compatibility
guided_options_request = {"guided_json": response_format["json_schema"]["schema"]} if response_format else None
structured_outputs = (
StructuredOutputsParams(json=response_format["json_schema"]["schema"]) if response_format else None
)

messages = completion_kwargs.pop("messages")
prepared_stop_sequences = completion_kwargs.pop("stop", [])
Expand All @@ -658,12 +661,12 @@ def generate(
temperature=kwargs.get("temperature", 0.0),
max_tokens=kwargs.get("max_tokens", 2048),
stop=prepared_stop_sequences,
structured_outputs=structured_outputs,
)

out = self.model.generate(
prompt,
sampling_params=sampling_params,
guided_options_request=guided_options_request,
**completion_kwargs,
)

Expand Down