add decode round robin policy (#15164)
This commit is contained in:
@@ -3136,6 +3136,7 @@ class ServerArgs:
|
||||
help="The load balancing strategy for data parallelism.",
|
||||
choices=[
|
||||
"round_robin",
|
||||
"decode_round_robin",
|
||||
"shortest_queue",
|
||||
"minimum_tokens",
|
||||
],
|
||||
|
||||
Reference in New Issue
Block a user