From 19df1aec2bd586e5273eee2b593c8280f4856522 Mon Sep 17 00:00:00 2001 From: xjkmfa <108254620+xjkmfa@users.noreply.github.com> Date: Wed, 29 Oct 2025 17:39:40 +0800 Subject: [PATCH] [Docs] add Qwen25vl yaml (#4662) MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit * Add ci case for min token and max token * 【CI case】include total_tokens in the last packet of completion interface stream output * 【CE】add qwen25-vl * 【CE】add qwen25-vl --------- Co-authored-by: xujing43 --- benchmarks/yaml/qwen25_7b-vl-32k-bf16.yaml | 6 ++++++ benchmarks/yaml/request_yaml/qwen25-vl-32k.yaml | 8 ++++++++ 2 files changed, 14 insertions(+) create mode 100644 benchmarks/yaml/qwen25_7b-vl-32k-bf16.yaml create mode 100644 benchmarks/yaml/request_yaml/qwen25-vl-32k.yaml diff --git a/benchmarks/yaml/qwen25_7b-vl-32k-bf16.yaml b/benchmarks/yaml/qwen25_7b-vl-32k-bf16.yaml new file mode 100644 index 000000000..d159e676f --- /dev/null +++ b/benchmarks/yaml/qwen25_7b-vl-32k-bf16.yaml @@ -0,0 +1,6 @@ +max_model_len: 32768 +max_num_seqs: 128 +gpu_memory_utilization: 0.85 +tensor_parallel_size: 1 +limit_mm_per_prompt: '{"image": 100, "video": 100}' +enable_mm: True \ No newline at end of file diff --git a/benchmarks/yaml/request_yaml/qwen25-vl-32k.yaml b/benchmarks/yaml/request_yaml/qwen25-vl-32k.yaml new file mode 100644 index 000000000..0c9a944e6 --- /dev/null +++ b/benchmarks/yaml/request_yaml/qwen25-vl-32k.yaml @@ -0,0 +1,8 @@ +top_p: 0.8 +temperature: 0.7 +metadata: + min_tokens: 1 +max_tokens: 32768 +repetition_penalty: 1.05 +frequency_penalty: 0 +presence_penalty: 0 \ No newline at end of file