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 12, 17:15 UTC
[Frontend][Responses API] Multi-turn (with type: "output_text") support for non-harmony requests (#29175)
Signed-off-by: Mads Kildegård <mkildegaard99@gmail.com>
Mads Kildegård committed
Nov 22, 2025 at 10:58 UTC
ea38474ac564efdc09762ad066139b75cf68f924
1 file changed
+2
-1
vllm/entrypoints/chat_utils.py
+2
-1
@@ -1283,6 +1283,7 @@ MM_PARSER_MAP: dict[
1283
"text": lambda part: _TextParser(part).get("text", None),
1284
"thinking": lambda part: _ThinkParser(part).get("thinking", None),
1285
"input_text": lambda part: _TextParser(part).get("text", None),
1286
+ "output_text": lambda part: _TextParser(part).get("text", None),
1287
"input_image": lambda part: _ResponsesInputImageParser(part).get("image_url", None),
1288
"image_url": lambda part: _ImageParser(part).get("image_url", {}).get("url", None),
1289
"image_embeds": lambda part: _ImageEmbedsParser(part).get("image_embeds", None),
@@ -1463,7 +1464,7 @@ def _parse_chat_message_content_part(
1464
)
1465
return None
1466
1466
- if part_type in ("text", "input_text", "refusal", "thinking"):
1467
+ if part_type in ("text", "input_text", "output_text", "refusal", "thinking"):
1468
str_content = cast(str, content)
1469
if wrap_dicts:
1470
return {"type": "text", "text": str_content}