A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 4, 07:30 UTC
[Benchmark] Add --trust-remote-code flag to multi-turn benchmark (#41661)
Signed-off-by: Dao Le <daole@inferact.ai> Signed-off-by: Dao Le <Dao007forever@gmail.com> Co-authored-by: Claude <noreply@anthropic.com>
Dao007forever committed
May 5, 2026 at 01:00 UTC
4845aee6b78c5543e036df1f9bbb8a9bfa4961f2
1 file changed
+9
-1
benchmarks/multi_turn/benchmark_serving_multi_turn.py
+9
-1
@@ -1473,6 +1473,12 @@ async def main() -> None:
1473
"(for example: --warmup-percentages=0%%,50%%)",
1474
)
1475
1476
+ parser.add_argument(
1477
+ "--trust-remote-code",
1478
+ action="store_true",
1479
+ help="Trust remote code when loading the tokenizer.",
1480
+ )
1481
+
1482
args = parser.parse_args()
1483
1484
logger.info(args)
@@ -1515,7 +1521,9 @@ async def main() -> None:
1521
np.random.seed(args.seed)
1522
1523
logger.info("Loading tokenizer")
1518
- tokenizer = AutoTokenizer.from_pretrained(args.model)
1524
+ tokenizer = AutoTokenizer.from_pretrained(
1525
+ args.model, trust_remote_code=args.trust_remote_code
1526
+ )
1527
1528
await get_server_info(args.url)
1529