Support chunked prefill when radix cache is disabled (#811)

2024-08-01 00:29:01 -07:00
parent ca600e8cd6
commit c020f9ceda
9 changed files with 163 additions and 26 deletions
--- a/python/sglang/srt/server_args.py
+++ b/python/sglang/srt/server_args.py
@@ -419,10 +419,6 @@ class ServerArgs:
            self.dp_size > 1 and self.node_rank is not None
        ), "multi-node data parallel is not supported"

-        assert not (
-            self.chunked_prefill_size is not None and self.disable_radix_cache
-        ), "chunked prefill is not supported with radix cache disabled currently"
-

@dataclasses.dataclass
 class PortArgs: