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 15, 03:30 UTC
[Misc] remove redundant char (#20287)
Signed-off-by: Kebe <mail@kebe7jun.com>
Kebe committed
Jul 1, 2025 at 15:33 UTC
b1c1fe35a599cfd3c0404702c65c2381b025bc6a
2 files changed
+2
-2
benchmarks/benchmark_serving.py
+1
-1
@@ -551,7 +551,7 @@ async def benchmark(
551
"total_input_tokens": metrics.total_input,
552
"total_output_tokens": metrics.total_output,
553
"request_throughput": metrics.request_throughput,
554
- "request_goodput:": metrics.request_goodput if goodput_config_dict else None,
554
+ "request_goodput": metrics.request_goodput if goodput_config_dict else None,
555
"output_throughput": metrics.output_throughput,
556
"total_token_throughput": metrics.total_token_throughput,
557
"input_lens": [output.prompt_len for output in outputs],
vllm/benchmarks/serve.py
+1
-1
@@ -498,7 +498,7 @@ async def benchmark(
498
"total_input_tokens": metrics.total_input,
499
"total_output_tokens": metrics.total_output,
500
"request_throughput": metrics.request_throughput,
501
- "request_goodput:":
501
+ "request_goodput":
502
metrics.request_goodput if goodput_config_dict else None,
503
"output_throughput": metrics.output_throughput,
504
"total_token_throughput": metrics.total_token_throughput,