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 11, 08:45 UTC
[CI/Build] Fix compatibility between #30244 and #30396 (#30787)
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
Cyrus Leung committed
Dec 17, 2025 at 12:21 UTC
44d3b1df3d6416b76d84c360d751b8f5220c0b11
1 file changed
+3
-1
tests/compile/distributed/test_fusions_e2e.py
+3
-1
@@ -564,7 +564,9 @@ def test_rms_group_quant(
564
splitting_ops=splitting_ops,
565
# Common
566
mode=CompilationMode.VLLM_COMPILE,
567
- pass_config=PassConfig(eliminate_noops=True, fuse_norm_quant=True),
567
+ pass_config=PassConfig(
568
+ fuse_norm_quant=True, fuse_act_quant=True, eliminate_noops=True
569
+ ),
570
# Inductor caches custom passes by default as well via uuid
571
inductor_compile_config={"force_disable_caches": True},
572
)