xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2026-09-10 10:27:04 +08:00
Code Issues Packages Projects Releases Wiki Activity

misc cleanups to prepare for rebase

Browse Source 
Signed-off-by: Sage Moore <sage@neuralmagic.com>
This commit is contained in:
Sage Moore 2025-06-02 14:15:52 +00:00
parent 0323e29153
commit 8f592524cb
2 changed files with 1 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
vllm/envs.py
View File

@ -811,10 +811,6 @@ environment_variables: dict[str, Callable[[], Any]] = {
# all2all backend for vllm's expert parallel communication # all2all backend for vllm's expert parallel communication
"VLLM_ALL2ALL_BACKEND": "VLLM_ALL2ALL_BACKEND":
lambda: os.getenv("VLLM_ALL2ALL_BACKEND", "naive"), lambda: os.getenv("VLLM_ALL2ALL_BACKEND", "naive"),
# check that the cudagraphs input addresses are correct before replaying
"VLLM_CUDAGRAPH_SANITIZER":
lambda: os.getenv("VLLM_CUDAGRAPH_SANITIZER", "0") == "1",
} }
# end-env-vars-definition # end-env-vars-definition

2
vllm/model_executor/layers/fused_moe/layer.py
View File

@ -29,7 +29,6 @@ from vllm.model_executor.utils import set_weight_attrs
from vllm.platforms import current_platform from vllm.platforms import current_platform
from vllm.platforms.interface import CpuArchEnum from vllm.platforms.interface import CpuArchEnum
from vllm.utils import direct_register_custom_op from vllm.utils import direct_register_custom_op
from vllm.v1.worker.ubatching import get_current_ubatch_context
has_pplx = importlib.util.find_spec("pplx_kernels") is not None has_pplx = importlib.util.find_spec("pplx_kernels") is not None
@ -306,6 +305,7 @@ class AllToAllCache:
self._cache[key] = instance self._cache[key] = instance
return instance return instance
# Global singleton # Global singleton
_all_to_all_cache = AllToAllCache() _all_to_all_cache = AllToAllCache()