Skip to content

Commit 5fbe593

Browse files
Merge pull request #447 from hodaaaaaaaaaa/chore/vllm-no-usage-stats
Disable vLLM usage reporting in the sampler
2 parents a600114 + 421866c commit 5fbe593

1 file changed

Lines changed: 3 additions & 0 deletions

File tree

‎src/server/vllm_sampler.py‎

Lines changed: 3 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -174,6 +174,9 @@ def make_engine(sleep_mode: bool, is_lora: bool) -> Any:
174174
from vllm.engine.arg_utils import AsyncEngineArgs
175175
from vllm.engine.async_llm_engine import AsyncLLMEngine
176176

177+
# vLLM posts hardware, model and config details to stats.vllm.ai unless told not to; the engine subprocess
178+
# inherits this. Operators can set the variable to 0 to opt back in.
179+
os.environ.setdefault("VLLM_NO_USAGE_STATS", "1")
177180
args = engine_args(is_lora, sleep_mode=sleep_mode)
178181
if transfer := args.get("weight_transfer_config"):
179182
args["weight_transfer_config"] = WeightTransferConfig(**transfer)

0 commit comments

Comments
 (0)