A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 17, 03:00 UTC
[Bugfix] Remove key sorting for `guided_json` parameter in OpenAi compatible Server (#3945)
Daniel E Marasco committed
Apr 10, 2024 at 13:15 UTC
e4c4072c94b346053768691451566c56664e26a7
1 file changed
+1
-1
vllm/model_executor/guided_decoding.py
+1
-1
@@ -91,7 +91,7 @@ def _get_guide_and_mode(
91
json = request.guided_json
92
if isinstance(json, dict):
93
# turn dict into hashable string
94
- json = json_dumps(json, sort_keys=True)
94
+ json = json_dumps(json)
95
elif isinstance(json, BaseModel):
96
# use pydantic signature so that different model classes
97
# with the same fields will get hashed the same