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 2, 08:30 UTC
[BugFix][Platform] Fix import vllm.platforms.rocm error on non-CUDA test_gpt_oss.py (#43571)
Signed-off-by: Ma, Liangliang <liangliang.ma@intel.com> Co-authored-by: Kunshang Ji <kunshang.ji@intel.com>
Liangliang Ma committed
May 30, 2026 at 14:16 UTC
e9499996df8968f473db1f6bc7ec31207022aea0
1 file changed
+8
-1
tests/models/quantization/test_gpt_oss.py
+8
-1
@@ -22,7 +22,14 @@ import pytest
22
from packaging import version
23
24
from vllm.platforms import current_platform
25
-from vllm.platforms.rocm import on_gfx950
25
+
26
+if current_platform.is_rocm():
27
+ from vllm.platforms.rocm import on_gfx950
28
+else:
29
+
30
+ def on_gfx950() -> bool:
31
+ return False
32
+
33
34
MODEL_ACCURACIES = {
35
# Full quantization: attention linears and MoE linears