From 613411a90ccfd884383aab84875ce95fe2871a3c Mon Sep 17 00:00:00 2001 From: Wentao Ye <44945378+yewentao256@users.noreply.github.com> Date: Wed, 5 Aug 2026 14:33:45 -0400 Subject: [PATCH] [CI Bug] Fix `Chunked prefill is required for mamba cache mode 'align'.` (#51177) Signed-off-by: yewentao256 --- tests/quantization/test_experts_int8.py | 1 + 1 file changed, 1 insertion(+) diff --git a/tests/quantization/test_experts_int8.py b/tests/quantization/test_experts_int8.py index 6119c1c846..d35cf9912a 100644 --- a/tests/quantization/test_experts_int8.py +++ b/tests/quantization/test_experts_int8.py @@ -38,5 +38,6 @@ def test_model_experts_int8_startup( dtype=dtype, enforce_eager=True, quantization="experts_int8", + enable_chunked_prefill=True, ) as vllm_model: vllm_model.generate_greedy(example_prompts, max_tokens)