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, 23:45 UTC
[Misc] fixed nvfp4_moe test failures due to invalid kwargs (#21246)
Signed-off-by: Yang Chen <yangche@fb.com>
Yang Chen committed
Jul 23, 2025 at 01:41 UTC
6929f8b437d2c6a3c8ff244d439fda36f7dab0ca
1 file changed
+2
-2
tests/kernels/moe/test_nvfp4_moe.py
+2
-2
@@ -93,11 +93,11 @@ def test_cutlass_fp4_moe_no_graph(m: int, n: int, k: int, e: int, topk: int,
93
a1_gscale=a1_gs,
94
w1_fp4=w1_q,
95
w1_blockscale=w1_blockscale,
96
- w1_alphas=(1 / w1_gs),
96
+ g1_alphas=(1 / w1_gs),
97
a2_gscale=a2_gs,
98
w2_fp4=w2_q,
99
w2_blockscale=w2_blockscale,
100
- w2_alphas=(1 / w2_gs),
100
+ g2_alphas=(1 / w2_gs),
101
topk_weights=topk_weights,
102
topk_ids=topk_ids,
103
m=m,