From 0414d7541033596894e72ca75cf29df593ad60fc Mon Sep 17 00:00:00 2001 From: Yan Ma Date: Thu, 4 Jun 2026 08:48:17 +0800 Subject: [PATCH] [XPU] skip unapplied UT in test_gpu_model_runner.py (#44289) Signed-off-by: Yan Ma Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com> --- tests/v1/worker/test_gpu_model_runner.py | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/tests/v1/worker/test_gpu_model_runner.py b/tests/v1/worker/test_gpu_model_runner.py index 1a1352249c3..9642bfd79f8 100644 --- a/tests/v1/worker/test_gpu_model_runner.py +++ b/tests/v1/worker/test_gpu_model_runner.py @@ -1072,8 +1072,8 @@ def test_init_kv_cache_with_kv_sharing_valid(default_vllm_config): @pytest.mark.skipif( - current_platform.is_rocm(), - reason="Attention backend FLASHINFER is not supported on ROCm.", + not current_platform.is_cuda(), + reason="Attention backend FLASHINFER is only supported on CUDA.", ) def test_hybrid_attention_mamba_tensor_shapes(): """ @@ -1508,8 +1508,8 @@ def test_is_uniform_decode() -> None: @pytest.mark.skipif( - current_platform.is_rocm(), - reason="Attention backend FLASHINFER is not supported on ROCm.", + not current_platform.is_cuda(), + reason="Attention backend FLASHINFER is only supported on CUDA.", ) def test_mamba_cache_raises_when_max_num_seqs_exceeds_blocks(): """Test that a ValueError is raised when max_num_seqs exceeds the