phoenix-oss/llama-stack-mirror
0
0
Fork 1
mirror of https://github.com/meta-llama/llama-stack.git synced 2025-08-01 16:24:44 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

CR fixes.

Browse source
This commit is contained in:
ilya-kolchinsky 2025-05-12 20:55:24 +02:00
parent 3be85c717f
commit 89ed374fe3
1 changed files with 6 additions and 11 deletions
Download patch file Download diff file Expand all files Collapse all files

17
llama_stack/providers/remote/inference/vllm/vllm.py
View file

@ -167,23 +167,16 @@ def _process_vllm_chat_completion_end_of_stream(
chunks = []
if finish_reason is not None:
actual_finish_reason = _convert_to_vllm_finish_reason(finish_reason)
stop_reason = _convert_to_vllm_finish_reason(finish_reason)
else:
actual_finish_reason = StopReason.end_of_message
stop_reason = StopReason.end_of_message
if tool_call_buf.tool_name:
# at least one tool call request is received
args_str = tool_call_buf.arguments or "{}"
args = {}
args_parsed_successfully = True
try:
args = json.loads(args_str)
except Exception as e:
args_parsed_successfully = False
log.warning(f"Failed to parse tool call buffer arguments: {args_str} \nError: {e}")
if args_parsed_successfully:
chunks.append(
ChatCompletionResponseStreamChunk(
event=ChatCompletionResponseEvent(
@ -200,7 +193,9 @@ def _process_vllm_chat_completion_end_of_stream(
)
)
)
else:
except Exception as e:
log.warning(f"Failed to parse tool call buffer arguments: {args_str} \nError: {e}")
chunks.append(
ChatCompletionResponseStreamChunk(
event=ChatCompletionResponseEvent(
@ -219,7 +214,7 @@ def _process_vllm_chat_completion_end_of_stream(
event_type=ChatCompletionResponseEventType.complete,
delta=TextDelta(text=last_chunk_content or ""),
logprobs=None,
stop_reason=actual_finish_reason,
stop_reason=stop_reason,
)
)
)