From 2c3ecf32f1a8be038712a1d663ced6eae722a1c1 Mon Sep 17 00:00:00 2001 From: Dustin Luong <24824135+dstnluong@users.noreply.github.com> Date: Thu, 6 Aug 2026 17:00:48 -0700 Subject: [PATCH] [Fix] Reformat /vertex_generate successful predictions (#33446) Co-authored-by: Jimmy Shong <69131491+Jiminator@users.noreply.github.com> --- python/sglang/srt/entrypoints/http_server.py | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/python/sglang/srt/entrypoints/http_server.py b/python/sglang/srt/entrypoints/http_server.py index ac38b4668..a2f48d9b7 100644 --- a/python/sglang/srt/entrypoints/http_server.py +++ b/python/sglang/srt/entrypoints/http_server.py @@ -2041,8 +2041,10 @@ async def vertex_generate( **(vertex_req.parameters or {}), ) ret = await generate_request(req, raw_request) - if isinstance(ret, Response): - return ret + if isinstance(ret, Response) and ret.status_code == 200 and hasattr(ret, "body"): + return ORJSONResponse({"predictions": orjson.loads(ret.body)}) + return ret + return ORJSONResponse({"predictions": ret})