mirror of
https://github.com/PaddlePaddle/FastDeploy.git
synced 2025-12-24 13:28:13 +08:00
[Optimization] Improve perf for fd response token with internal adapter (#4992)
* [Optimize] Improve perf for fd response token with internal adapter * fix * fix bug * fix ci * fix ci * fix ci * fix ci
This commit is contained in:
@@ -105,11 +105,15 @@ environment_variables: dict[str, Callable[[], Any]] = {
|
||||
"FD_ENABLE_MODEL_CACHE": lambda: bool(int(os.getenv("FD_ENABLE_MODEL_CACHE", "0"))),
|
||||
# enable internal module to access LLMEngine.
|
||||
"FD_ENABLE_INTERNAL_ADAPTER": lambda: int(os.getenv("FD_ENABLE_INTERNAL_ADAPTER", "0")),
|
||||
# LLMEngine recieve requests port, used when FD_ENABLE_INTERNAL_ADAPTER=1
|
||||
# LLMEngine receive requests port, used when FD_ENABLE_INTERNAL_ADAPTER=1
|
||||
"FD_ZMQ_RECV_REQUEST_SERVER_PORT": lambda: os.getenv("FD_ZMQ_RECV_REQUEST_SERVER_PORT", "8200"),
|
||||
# LLMEngine send response port, used when FD_ENABLE_INTERNAL_ADAPTER=1
|
||||
"FD_ZMQ_SEND_RESPONSE_SERVER_PORT": lambda: os.getenv("FD_ZMQ_SEND_RESPONSE_SERVER_PORT", "8201"),
|
||||
# LLMEngine recieve control command port, used when FD_ENABLE_INTERNAL_ADAPTER=1
|
||||
# LLMEngine receive requests port, used when FD_ENABLE_INTERNAL_ADAPTER=1
|
||||
"FD_ZMQ_RECV_REQUEST_SERVER_PORTS": lambda: os.getenv("FD_ZMQ_RECV_REQUEST_SERVER_PORTS", "8200"),
|
||||
# LLMEngine send response port, used when FD_ENABLE_INTERNAL_ADAPTER=1
|
||||
"FD_ZMQ_SEND_RESPONSE_SERVER_PORTS": lambda: os.getenv("FD_ZMQ_SEND_RESPONSE_SERVER_PORTS", "8201"),
|
||||
# LLMEngine receive control command port, used when FD_ENABLE_INTERNAL_ADAPTER=1
|
||||
"FD_ZMQ_CONTROL_CMD_SERVER_PORTS": lambda: os.getenv("FD_ZMQ_CONTROL_CMD_SERVER_PORTS", "8202"),
|
||||
# Whether to enable cache task in decode node
|
||||
"FD_ENABLE_CACHE_TASK": lambda: os.getenv("FD_ENABLE_CACHE_TASK", "1"),
|
||||
|
||||
Reference in New Issue
Block a user