A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 14, 05:30 UTC
[CI] Temporarily disable flaky test (#22930)
Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
Lucas Wilkinson committed
Aug 14, 2025 at 15:59 UTC
b8ff05361a2ab91e6be33601d4f564408e10eb24
1 file changed
+6
tests/v1/e2e/test_spec_decode.py
+6
@@ -162,6 +162,12 @@ def test_eagle_correctness(
162
mm_enabled: bool,
163
attn_backend: str,
164
):
165
+ if attn_backend == "TREE_ATTN":
166
+ # TODO: Fix this flaky test
167
+ pytest.skip(
168
+ "TREE_ATTN is flaky in the test disable for now until it can be "
169
+ "reolved (see https://github.com/vllm-project/vllm/issues/22922)")
170
+
171
# Generate test prompts inside the function instead of using fixture
172
test_prompts = get_test_prompts(mm_enabled)
173
'''