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 8, 10:00 UTC
Fix CI failure - Flashinfer Kernel tests (#34316)
Signed-off-by: wzhao18 <wzhao18.sz@gmail.com>
Wei Zhao committed
Feb 11, 2026 at 17:17 UTC
5aff2699bdcedd9ee91fe936fc21b26466203ae1
3 files changed
+3
tests/kernels/moe/test_flashinfer.py
+1
@@ -287,6 +287,7 @@ def test_flashinfer_cutlass_moe_fp8_no_graph(
287
hidden_dim=k,
288
intermediate_size_per_partition=n,
289
num_local_experts=e,
290
+ num_logical_experts=e,
291
activation=activation,
292
device="cuda",
293
moe_parallel_config=FusedMoEParallelConfig.make_no_parallel(),
tests/kernels/moe/test_flashinfer_moe.py
+1
@@ -97,6 +97,7 @@ def test_flashinfer_fp4_moe_no_graph(
97
hidden_dim=k,
98
intermediate_size_per_partition=n,
99
num_local_experts=e,
100
+ num_logical_experts=e,
101
activation=activation,
102
device="cuda",
103
moe_parallel_config=FusedMoEParallelConfig.make_no_parallel(),
tests/kernels/moe/test_pplx_cutlass_moe.py
+1
@@ -147,6 +147,7 @@ def pplx_cutlass_moe(
147
hidden_dim=hidden_dim,
148
intermediate_size_per_partition=intermediate_dim,
149
num_local_experts=num_local_experts,
150
+ num_logical_experts=num_experts,
151
moe_parallel_config=FusedMoEParallelConfig.make_no_parallel(),
152
activation="silu",
153
in_dtype=torch.bfloat16,