feat: support flashinfer mla with prefix cache (#3643)

2025-02-18 02:06:43 +08:00
parent c38f3aed24
commit 714f3e6362
4 changed files with 107 additions and 31 deletions
--- a/python/sglang/srt/managers/schedule_batch.py
+++ b/python/sglang/srt/managers/schedule_batch.py
@@ -66,6 +66,7 @@ global_server_args_dict = {
    "enable_ep_moe": ServerArgs.enable_ep_moe,
    "device": ServerArgs.device,
    "enable_flashinfer_mla": ServerArgs.enable_flashinfer_mla,
+    "disable_radix_cache": ServerArgs.disable_radix_cache,
 }

 logger = logging.getLogger(__name__)