diff --git a/benchmarks/benchmark_throughput.py b/benchmarks/benchmark_throughput.py index c5526c1c40ea..39629d08d639 100644 --- a/benchmarks/benchmark_throughput.py +++ b/benchmarks/benchmark_throughput.py @@ -97,6 +97,7 @@ def run_cacheflow( # FIXME(woosuk): Do not use internal method. llm._add_request( prompt=prompt, + prompt_token_ids=None, sampling_params=sampling_params, )