From 185bc91d79d64ca8b52c6fe2c2b0e98878c9e8f6 Mon Sep 17 00:00:00 2001 From: "Li, Tianmu" Date: Mon, 22 Jun 2026 18:18:11 +0000 Subject: [PATCH 1/2] Add qwen perf yaml Signed-off-by: Li, Tianmu --- .../qwen_agentic_benchmark.yaml | 40 +++++++++++++++++++ 1 file changed, 40 insertions(+) create mode 100644 examples/10_Agentic_Inference/qwen_agentic_benchmark.yaml diff --git a/examples/10_Agentic_Inference/qwen_agentic_benchmark.yaml b/examples/10_Agentic_Inference/qwen_agentic_benchmark.yaml new file mode 100644 index 000000000..b3d4ada7e --- /dev/null +++ b/examples/10_Agentic_Inference/qwen_agentic_benchmark.yaml @@ -0,0 +1,40 @@ +name: "qwen-agentic-benchmark" +version: "1.0" +type: "online" + +model_params: + name: "Qwen/Qwen3.6-35B-A3B" + temperature: 1.0 + top_k: 20 + top_p: 0.95 + repetition_penalty: 1.0 + presence_penalty: 1.5 + max_new_tokens: 8192 + chat_template_kwargs: + preserve_thinking: true + +datasets: + - name: agentic_coding + type: performance + path: /path/to/agentic_combined.jsonl + accuracy_config: + eval_method: agentic_inference_inline # required benchmark default. + agentic_inference: + enable_salt: true # do not change. + inject_tool_delay: true # do not change. + +settings: + runtime: + min_duration_ms: 0 + max_duration_ms: 36000000 + + load_pattern: + type: agentic_inference + target_concurrency: 8 # Submission-specific concurrency. + +endpoint_config: + endpoints: + - "http://localhost:30000" + api_type: openai + +report_dir: logs/qwen_agentic From b5ea6c281d67a4775acdf427e83592bcbc785bd2 Mon Sep 17 00:00:00 2001 From: "Li, Tianmu" Date: Mon, 22 Jun 2026 18:19:41 +0000 Subject: [PATCH 2/2] Minor typo Signed-off-by: Li, Tianmu --- examples/10_Agentic_Inference/qwen_agentic_benchmark.yaml | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/examples/10_Agentic_Inference/qwen_agentic_benchmark.yaml b/examples/10_Agentic_Inference/qwen_agentic_benchmark.yaml index b3d4ada7e..7e131cbe2 100644 --- a/examples/10_Agentic_Inference/qwen_agentic_benchmark.yaml +++ b/examples/10_Agentic_Inference/qwen_agentic_benchmark.yaml @@ -14,7 +14,7 @@ model_params: preserve_thinking: true datasets: - - name: agentic_coding + - name: agentic_combined type: performance path: /path/to/agentic_combined.jsonl accuracy_config: