apps/FastDeploy
1
0
Fork 0
mirror of https://github.com/PaddlePaddle/FastDeploy.git synced 2025-10-01 06:42:23 +08:00
Code Issues Actions 4 Packages Projects Releases Wiki Activity

[BugFix] fix parameter is 0 (#3592)

Browse Source 
* Update engine_client.py

* fix

* Update common_engine.py
This commit is contained in:
ltd0924
2025-08-28 09:52:36 +08:00
committed by GitHub
parent 479c8b85d3
commit 3d92fb09f7
3 changed files with 28 additions and 25 deletions
Download Patch File Download Diff File Expand all files Collapse all files

34
fastdeploy/engine/common_engine.py
View File

@@ -592,18 +592,23 @@ class EngineSevice:
request, insert_task = None, []
results: List[Tuple[str, Optional[str]]] = list()
if data:
request = Request.from_dict(data)
start_span("ENQUEUE_ZMQ", data, trace.SpanKind.PRODUCER)
llm_logger.debug(f"Receive request: {request}")
err_msg = None
if self.guided_decoding_checker is not None:
request, err_msg = self.guided_decoding_checker.schema_format(request)
try:
request = Request.from_dict(data)
start_span("ENQUEUE_ZMQ", data, trace.SpanKind.PRODUCER)
llm_logger.debug(f"Receive request: {request}")
except Exception as e:
llm_logger.error(f"Receive request error: {e}, {traceback.format_exc()!s}")
err_msg = str(e)
results.append((data["request_id"], err_msg))
if err_msg is not None:
llm_logger.error(err_msg)
results.append((request.request_id, err_msg))
else:
if self.guided_decoding_checker is not None and err_msg is None:
request, err_msg = self.guided_decoding_checker.schema_format(request)
if err_msg is not None:
llm_logger.error(f"Receive request error: {err_msg}")
results.append((request.request_id, err_msg))
if err_msg is None:
insert_task.append(request)
response = self.scheduler.put_requests(insert_task)
@@ -615,9 +620,10 @@ class EngineSevice:
added_requests[request.request_id] += 1
for request_id, failed in results:
added_requests[request_id] -= 1
if added_requests[request_id] == 0:
added_requests.pop(request_id)
if request_id in added_requests:
added_requests[request_id] -= 1
if added_requests[request_id] == 0:
added_requests.pop(request_id)
if failed is None:
main_process_metrics.num_requests_waiting.inc(1)
@@ -631,7 +637,7 @@ class EngineSevice:
)
# Since the request is not in scheduler
# Send result by zmq directly
self.zmq_server.send_multipart(request_id, error_result)
self.zmq_server.send_multipart(request_id, [error_result])
except Exception as e:
llm_logger.error(
f"Error happend while receving new request from zmq, details={e}, "