squall/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity

Disable Ray usage stats collection (#2206)

Browse Source
This commit is contained in:
Woosuk Kwon 2023-12-20 21:52:08 -08:00 committed by GitHub
parent c17daa9f89
commit 3a4fd5ca59
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 5 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
vllm/engine/llm_engine.py
View File

@ -1,4 +1,5 @@
import copy import copy
import os
import time import time
from functools import partial from functools import partial
from typing import TYPE_CHECKING, Any, Iterable, List, Optional, Tuple, Union from typing import TYPE_CHECKING, Any, Iterable, List, Optional, Tuple, Union
@ -105,6 +106,10 @@ class LLMEngine:
# Create the parallel GPU workers. # Create the parallel GPU workers.
if self.parallel_config.worker_use_ray: if self.parallel_config.worker_use_ray:
# Disable Ray usage stats collection.
ray_usage = os.environ.get("RAY_USAGE_STATS_ENABLED", "0")
if ray_usage != "1":
os.environ["RAY_USAGE_STATS_ENABLED"] = "0"
self._init_workers_ray(placement_group) self._init_workers_ray(placement_group)
else: else:
self._init_workers(distributed_init_method) self._init_workers(distributed_init_method)