[Kernel][Misc] register ops to prevent graph breaks (#6917)
Co-authored-by: Sage Moore <sage@neuralmagic.com>
This commit is contained in:
@@ -203,6 +203,11 @@ environment_variables: Dict[str, Callable[[], Any]] = {
|
||||
(os.environ.get("VLLM_DYNAMO_USE_CUSTOM_DISPATCHER", "True").lower() in
|
||||
("true", "1")),
|
||||
|
||||
# Internal flag to enable Dynamo fullgraph capture
|
||||
"VLLM_TEST_DYNAMO_FULLGRAPH_CAPTURE":
|
||||
lambda: bool(
|
||||
os.environ.get("VLLM_TEST_DYNAMO_FULLGRAPH_CAPTURE", "1") != "0"),
|
||||
|
||||
# local rank of the process in the distributed setting, used to determine
|
||||
# the GPU device id
|
||||
"LOCAL_RANK":
|
||||
|
||||
Reference in New Issue
Block a user