apps/FastDeploy
1
0
Fork 0
mirror of https://github.com/PaddlePaddle/FastDeploy.git synced 2025-10-06 00:57:33 +08:00
Code Issues Actions 2 Packages Projects Releases Wiki Activity

[BUGFIX] clear request (#4320)
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

Browse Source 
Co-authored-by: ltd0924 <luotingdan@baidu.com>
This commit is contained in:
ltd0924
2025-09-29 20:37:58 +08:00
committed by GitHub
parent 63a03ee152
commit e42dc8c694
4 changed files with 6 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
fastdeploy/engine/common_engine.py
View File

@@ -806,7 +806,8 @@ class EngineSevice:
llm_logger.info("Clear Data: Start") llm_logger.info("Clear Data: Start")
self.token_processor.clear_data() self.token_processor.clear_data()
self.engine_worker_queue.clear_data() self.engine_worker_queue.clear_data()
self.zmq_server.req_dict.clear() self.send_response_server.req_dict.clear()
self.recv_request_server.req_dict.clear()
llm_logger.info("Clear Data: Successfully") llm_logger.info("Clear Data: Successfully")
return True return True
except Exception as e: except Exception as e:

1
fastdeploy/entrypoints/openai/api_server.py
View File

@@ -499,6 +499,7 @@ def control_scheduler(request: ControlSchedulerRequest):
return JSONResponse(content=content.model_dump(), status_code=500) return JSONResponse(content=content.model_dump(), status_code=500)
if request.reset: if request.reset:
llm_engine.engine.clear_data()
llm_engine.engine.scheduler.reset() llm_engine.engine.scheduler.reset()
if request.load_shards_num or request.reallocate_shard: if request.load_shards_num or request.reallocate_shard:

1
fastdeploy/rl/dynamic_weight_manager.py
View File

@@ -223,6 +223,7 @@ class DynamicWeightManager:
while model_weights_status.value[0] != ModelWeightsStatus.NORMAL: while model_weights_status.value[0] != ModelWeightsStatus.NORMAL:
if model_weights_status.value[0] == ModelWeightsStatus.UPDATING: if model_weights_status.value[0] == ModelWeightsStatus.UPDATING:
logger.info("infer engine stopped! start to load new checkpoint...") logger.info("infer engine stopped! start to load new checkpoint...")
model_runner.clear_requests()
model_runner.update_parameters(pid) model_runner.update_parameters(pid)
while model_weights_status.value[0] != ModelWeightsStatus.NORMAL: while model_weights_status.value[0] != ModelWeightsStatus.NORMAL:
time.sleep(0.01) time.sleep(0.01)

2
fastdeploy/worker/worker_process.py
View File

@@ -337,6 +337,8 @@ class PaddleDisWorkerProc:
self.worker.model_runner, self.worker.model_runner,
self.parallel_config.engine_worker_queue_port, self.parallel_config.engine_worker_queue_port,
) )
logger.info(f"current task queue data: {self.task_queue.num_tasks()}")
self.task_queue.clear_data()
self.model_weights_signal[0] = ModelWeightsStatus.NORMAL self.model_weights_signal[0] = ModelWeightsStatus.NORMAL
logger.info(f"Rank: {self.local_rank} has updated or cleared parameters.") logger.info(f"Rank: {self.local_rank} has updated or cleared parameters.")