[MoE] Remove inplace fused experts mechanism (#43727)
Signed-off-by: Yongye Zhu <zyy1102000@gmail.com>
This commit is contained in:
@@ -233,7 +233,6 @@ def test_flashinfer_per_tensor_moe_fp8_no_graph(
|
||||
td.w2_quantized,
|
||||
topk_weights=topk_weights,
|
||||
topk_ids=topk_ids,
|
||||
inplace=False,
|
||||
activation=activation,
|
||||
global_num_experts=e,
|
||||
expert_map=None,
|
||||
@@ -321,7 +320,6 @@ def test_flashinfer_cutlass_moe_fp8_no_graph(
|
||||
td.w2_quantized,
|
||||
topk_weights=topk_weights,
|
||||
topk_ids=topk_ids,
|
||||
inplace=False,
|
||||
activation=activation,
|
||||
global_num_experts=e,
|
||||
expert_map=None,
|
||||
@@ -364,7 +362,6 @@ def test_flashinfer_cutlass_moe_fp8_no_graph(
|
||||
moe_config=moe_config,
|
||||
quant_config=quant_config,
|
||||
),
|
||||
inplace=False,
|
||||
)
|
||||
|
||||
flashinfer_cutlass_output = kernel.apply(
|
||||
|
||||
Reference in New Issue
Block a user