mirror of
https://github.com/meta-llama/llama-stack.git
synced 2025-08-02 08:44:44 +00:00
CR fixes.
This commit is contained in:
parent
3be85c717f
commit
89ed374fe3
1 changed files with 6 additions and 11 deletions
|
@ -167,23 +167,16 @@ def _process_vllm_chat_completion_end_of_stream(
|
||||||
chunks = []
|
chunks = []
|
||||||
|
|
||||||
if finish_reason is not None:
|
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:
|
else:
|
||||||
actual_finish_reason = StopReason.end_of_message
|
stop_reason = StopReason.end_of_message
|
||||||
|
|
||||||
if tool_call_buf.tool_name:
|
if tool_call_buf.tool_name:
|
||||||
# at least one tool call request is received
|
# at least one tool call request is received
|
||||||
|
|
||||||
args_str = tool_call_buf.arguments or "{}"
|
args_str = tool_call_buf.arguments or "{}"
|
||||||
args = {}
|
|
||||||
args_parsed_successfully = True
|
|
||||||
try:
|
try:
|
||||||
args = json.loads(args_str)
|
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(
|
chunks.append(
|
||||||
ChatCompletionResponseStreamChunk(
|
ChatCompletionResponseStreamChunk(
|
||||||
event=ChatCompletionResponseEvent(
|
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(
|
chunks.append(
|
||||||
ChatCompletionResponseStreamChunk(
|
ChatCompletionResponseStreamChunk(
|
||||||
event=ChatCompletionResponseEvent(
|
event=ChatCompletionResponseEvent(
|
||||||
|
@ -219,7 +214,7 @@ def _process_vllm_chat_completion_end_of_stream(
|
||||||
event_type=ChatCompletionResponseEventType.complete,
|
event_type=ChatCompletionResponseEventType.complete,
|
||||||
delta=TextDelta(text=last_chunk_content or ""),
|
delta=TextDelta(text=last_chunk_content or ""),
|
||||||
logprobs=None,
|
logprobs=None,
|
||||||
stop_reason=actual_finish_reason,
|
stop_reason=stop_reason,
|
||||||
)
|
)
|
||||||
)
|
)
|
||||||
)
|
)
|
||||||
|
|
Loading…
Add table
Add a link
Reference in a new issue