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 12, 02:15 UTC
[CI][AMD] Match Main CI Behavior By Skipping test_eplb_spec_decode In AMD CI (#30006)
Signed-off-by: Micah Williamson <micah.williamson@amd.com>
Micah Williamson committed
Dec 4, 2025 at 02:20 UTC
5430e110c099fdc6c8c80f443bf5adffe67aa30b
1 file changed
+7
tests/distributed/test_eplb_spec_decode.py
+7
@@ -6,6 +6,7 @@ import lm_eval
6
import pytest
7
8
from tests.utils import large_gpu_mark
9
+from vllm.platforms import current_platform
10
11
12
def get_model_args(
@@ -45,6 +46,12 @@ def get_model_args(
46
return model_args
47
48
49
+pytestmark = pytest.mark.skipif(
50
+ current_platform.is_rocm(),
51
+ reason="EPLB with Spec Decode is a work in progress on ROCm.",
52
+)
53
+
54
+
55
@pytest.mark.parametrize(
56
"model_setup",
57
[