biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[Kernel] Delegate construction of FusedMoEQuantConfig to FusedMoEMethodBase subclasses (#22537)

Browse Source 
Signed-off-by: Bill Nell <bnell@redhat.com>
This commit is contained in:
bnellnm
2025-09-17 19:43:31 -04:00
committed by GitHub
parent e6585ddb45
commit 5963b98b46
68 changed files with 2698 additions and 2526 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
tests/kernels/moe/test_batched_moe.py
View File

@@ -140,7 +140,7 @@ def test_batched_mm(num_experts: int, max_tokens_per_expert: int, K: int,
in_dtype=act_dtype,
quant_dtype=quant_dtype,
block_shape=block_shape,
per_act_token_quant=per_act_token_quant,
per_out_ch_quant=per_act_token_quant,
)
out_shape = (num_experts, max_tokens_per_expert, N)
@@ -250,7 +250,7 @@ def test_fused_moe_batched_experts(
block_shape=block_shape,
in_dtype=act_dtype,
quant_dtype=quant_dtype,
per_act_token_quant=per_act_token_quant,
per_out_ch_quant=per_act_token_quant,
)
if input_scales and quant_dtype is not None: