From 134ae2fb7b8746e8f1ca77a7a656e5fe1f796045 Mon Sep 17 00:00:00 2001 From: Sage Moore Date: Fri, 6 Dec 2024 15:59:39 +0000 Subject: [PATCH] format --- benchmarks/benchmark_latency.py | 7 +------ 1 file changed, 1 insertion(+), 6 deletions(-) diff --git a/benchmarks/benchmark_latency.py b/benchmarks/benchmark_latency.py index 23c683c6c37b5..50dea6c17c45b 100644 --- a/benchmarks/benchmark_latency.py +++ b/benchmarks/benchmark_latency.py @@ -5,7 +5,7 @@ import time from pathlib import Path from typing import List, Optional -from vllm.config import CompilationConfig + import numpy as np import torch from tqdm import tqdm @@ -21,11 +21,6 @@ def main(args: argparse.Namespace): engine_args = EngineArgs.from_cli_args(args) - config = CompilationConfig( - level=3, - custom_ops=["+silu_and_mul"], - ) - engine_args.compilation_config = config # NOTE(woosuk): If the request cannot be processed in a single batch, # the engine will automatically process the request in multiple batches. llm = LLM(**dataclasses.asdict(engine_args))