biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[Benchmark] multi_turn: Report warmup-inclusive runtime (#28937)

Browse Source 
Signed-off-by: Ido Segev <idos@pliops.com>
This commit is contained in:
Ido Segev
2025-11-18 18:38:22 +02:00
committed by GitHub
parent f6aa122698
commit 49a986ecd4
2 changed files with 53 additions and 10 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
benchmarks/multi_turn/README.md
View File

@@ -55,6 +55,10 @@ output_num_chunks 166.0 99.01 11.80 79.00 90.00 98.00 108.75
----------------------------------------------------------------------------------------------------
```
If you run with `--warmup-step`, the summary will also include `warmup_runtime_sec`
and `total_runtime_incl_warmup_sec` (while `runtime_sec` continues to reflect the
benchmark-only runtime so the reported throughput stays comparable).
### JSON configuration file for synthetic conversations generation
The input flag `--input-file` is used to determine the input conversations for the benchmark.<br/>