[CI Bug] Fix `Chunked prefill is required for mamba cache mode 'align'.` (#51177)
Signed-off-by: yewentao256 <zhyanwentao@126.com>
Wentao Ye committed
Aug 5, 2026 at 14:33 UTC
613411a90ccfd884383aab84875ce95fe2871a3c
1 file changed
+1
tests/quantization/test_experts_int8.py
+1
@@ -38,5 +38,6 @@ def test_model_experts_int8_startup(
38
dtype=dtype,
39
enforce_eager=True,
40
quantization="experts_int8",
41
+ enable_chunked_prefill=True,
42
) as vllm_model:
43
vllm_model.generate_greedy(example_prompts, max_tokens)