mirror of
https://github.com/vllm-project/vllm.git
synced 2025-10-20 14:53:52 +08:00
@ -22,7 +22,7 @@ from vllm.compilation.partition_rules import (
|
||||
resolve_defined_ops,
|
||||
)
|
||||
from vllm.config import CompilationConfig, CUDAGraphMode, VllmConfig
|
||||
from vllm.distributed.parallel_state import is_global_first_rank
|
||||
from vllm.distributed.parallel_state import is_local_first_rank
|
||||
from vllm.logger import init_logger
|
||||
from vllm.platforms import current_platform
|
||||
from vllm.utils import is_torch_equal_or_newer
|
||||
@ -604,7 +604,7 @@ class VllmBackend:
|
||||
|
||||
disable_cache = envs.VLLM_DISABLE_COMPILE_CACHE
|
||||
|
||||
if is_global_first_rank():
|
||||
if is_local_first_rank():
|
||||
if disable_cache:
|
||||
logger.info_once("vLLM's torch.compile cache is disabled.")
|
||||
else:
|
||||
|
Reference in New Issue
Block a user