apps/FastDeploy
1
0
Fork 0
mirror of https://github.com/PaddlePaddle/FastDeploy.git synced 2025-12-24 13:28:13 +08:00
Code Issues Actions 2 Packages Projects Releases Wiki Activity

[Bug fix] Send first token in D instance (#5199)
Some checks failed
CE Compile Job / ce_job_pre_check (push) Has been cancelled
Details
CE Compile Job / print_ce_job_pre_check_outputs (push) Has been cancelled
Details
CE Compile Job / FD-Clone-Linux (push) Has been cancelled
Details
CE Compile Job / Show Code Archive Output (push) Has been cancelled
Details
CE Compile Job / BUILD_SM8090 (push) Has been cancelled
Details
CE Compile Job / BUILD_SM8689 (push) Has been cancelled
Details
CE Compile Job / CE_UPLOAD (push) Has been cancelled
Details
Deploy GitHub Pages / deploy (push) Has been cancelled
Details

Browse Source 
* [Bug fix] Send first token in D instance

* fix
This commit is contained in:
chenjian
2025-11-24 23:42:20 +08:00
committed by GitHub
parent 95b39317a9
commit 09b47c7111
1 changed files with 5 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
fastdeploy/engine/common_engine.py
View File

@@ -1125,7 +1125,7 @@ class EngineService:
# received the request sent by the client
waiting_request_outputs.append(req_output)
continue
req_output.finished = False
ready_request_outputs.append(req_output)
self.llm_logger.debug(f"there are enough resource for prefilled request: {req_output.request_id}")
@@ -1145,6 +1145,8 @@ class EngineService:
self.resource_manager.pre_recycle_resource(request_id)
if request_id in self.token_processor.tokens_counter:
del self.token_processor.tokens_counter[request_id]
req_output.finished = True
self.scheduler.put_results([req_output])
continue
if req_output.error_code != 200:
self.llm_logger.warning(
@@ -1156,6 +1158,8 @@ class EngineService:
self.scheduler.put_results([req_output])
continue
self.token_processor.tokens_counter[request_id] = 1
if envs.FD_ENABLE_INTERNAL_ADAPTER: # first token sent by D instance
self.scheduler.put_results([req_output])
self.resource_manager.add_prefilled_request(req_output)
self.llm_logger.debug(f"add prefilled request success, {request_id}")