diff --git a/vllm/distributed/device_communicators/custom_all_reduce.py b/vllm/distributed/device_communicators/custom_all_reduce.py index 479dc95a..6229f1d6 100644 --- a/vllm/distributed/device_communicators/custom_all_reduce.py +++ b/vllm/distributed/device_communicators/custom_all_reduce.py @@ -15,7 +15,7 @@ from vllm.platforms import current_platform from vllm.utils import cuda_device_count_stateless try: - assert ops.is_custom_op_supported("_C_custom_ar::meta_size") + ops.meta_size() custom_ar = True except Exception: # For AMD GPUs and CPUs