diff --git a/examples/nixl/toy_proxy_server.py b/examples/nixl/toy_proxy_server.py index c0a26e485..74c1dc6fd 100644 --- a/examples/nixl/toy_proxy_server.py +++ b/examples/nixl/toy_proxy_server.py @@ -172,6 +172,7 @@ async def send_request_to_service( "do_remote_decode": True, "do_remote_prefill": False, "remote_engine_id": None, + "remote_request_id": f"prefill-{request_id}", "remote_block_ids": None, "remote_host": None, "remote_port": None, diff --git a/vllm_gaudi/platform.py b/vllm_gaudi/platform.py index 0f12ce30b..2838d6b34 100644 --- a/vllm_gaudi/platform.py +++ b/vllm_gaudi/platform.py @@ -50,6 +50,7 @@ def get_attn_backend_cls( use_mla: bool, has_sink: bool, use_sparse: bool, + use_mm_prefix: bool, attn_type: str | None = None, ) -> str: if use_sparse: