minke.yu/sglang
1
0
Fork 0
Code Issues Pull Requests Actions 12 Packages Projects Releases Wiki Activity

[Docs] Fix formatting in Evaluating New Models with SGLang (#17376)

Browse Source
This commit is contained in:
zijiexia
2026-01-19 18:22:30 -08:00
committed by GitHub
parent 7dc3cbe7ca
commit 9f8b79f16f
1 changed files with 8 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files
docs/developer_guide/evaluating_new_models.md
+8 -4
View File
@@ -52,8 +52,10 @@ python -m sglang.test.run_eval \
--max-tokens 120000 \ --max-tokens 120000 \
--repeat 8 --repeat 8
``` ```
> [!TIP]
> For reasoning models, add `--thinking-mode <mode>` (e.g., `qwen3`, `deepseek-r1`, `deepseek-v3`). You may skip it if the model has forced thinking enabled. ```{tip}
For reasoning models, add `--thinking-mode <mode>` (e.g., `qwen3`, `deepseek-r1`, `deepseek-v3`). You may skip it if the model has forced thinking enabled.
```
**HumanEval** **HumanEval**
@@ -75,8 +77,10 @@ python benchmark/mmmu/bench_sglang.py \
--port 30000 \ --port 30000 \
--concurrency 64 --concurrency 64
``` ```
> [!TIP]
> You can set max tokens by passing `--extra-request-body '{"max_tokens": 4096}'`. ```{tip}
You can set max tokens by passing `--extra-request-body '{"max_tokens": 4096}'`.
```
For models capable of processing video, we recommend extending the evaluation to include `VideoMME`, `MVBench`, and other relevant benchmarks. For models capable of processing video, we recommend extending the evaluation to include `VideoMME`, `MVBench`, and other relevant benchmarks.