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

Bugfix: Cutlass FP8 FusedMoE bad scaling factors (#27255)

Browse Source 
Signed-off-by: Amir Klein <203507526+amirkl94@users.noreply.github.com>
Co-authored-by: Tyler Michael Smith <tyler@neuralmagic.com>
Co-authored-by: Michael Goin <mgoin64@gmail.com>
This commit is contained in:
amirkl94
2025-11-05 13:06:06 +02:00
committed by GitHub
parent b57789b62b
commit 6b7a81185d
4 changed files with 40 additions and 14 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
vllm/model_executor/layers/fused_moe/config.py
View File

@@ -463,6 +463,10 @@ def fp8_w8a8_moe_quant_config(
per_act_token_quant: bool = False,
per_out_ch_quant: bool = False,
block_shape: list[int] | None = None,
a1_gscale: torch.Tensor | None = None,
a2_gscale: torch.Tensor | None = None,
g1_alphas: torch.Tensor | None = None,
g2_alphas: torch.Tensor | None = None,
) -> FusedMoEQuantConfig:
"""
Construct a quant config for fp8 activations and fp8 weights.
@@ -470,9 +474,13 @@ def fp8_w8a8_moe_quant_config(
return FusedMoEQuantConfig.make(
torch.float8_e4m3fn,
w1_scale=w1_scale,
g1_alphas=g1_alphas,
w2_scale=w2_scale,
g2_alphas=g2_alphas,
a1_scale=a1_scale,
a1_gscale=a1_gscale,
a2_scale=a2_scale,
a2_gscale=a2_gscale,
per_act_token_quant=per_act_token_quant,
per_out_ch_quant=per_out_ch_quant,
block_shape=block_shape,

14
vllm/model_executor/layers/fused_moe/flashinfer_cutlass_prepare_finalize.py
View File

@@ -170,7 +170,7 @@ class FlashInferAllGatherMoEPrepareAndFinalize(FlashInferCutlassMoEPrepareAndFin
self._apply_router_weight_on_input(
a1, topk_weights, topk_ids, apply_router_weight_on_input
)
if not self.use_dp:
if not self.use_dp and quant_config.quant_dtype == "nvfp4":
return a1, None, None, topk_ids, topk_weights
a1q, a1q_scale = moe_kernel_quantize_input(
@@ -181,11 +181,13 @@ class FlashInferAllGatherMoEPrepareAndFinalize(FlashInferCutlassMoEPrepareAndFin
quant_config.block_shape,
is_fp4_scale_swizzled=not self.use_dp,
)
topk_weights, topk_ids, a1q, a1q_scale = get_dp_group().all_gatherv(
[topk_weights, topk_ids, a1q, a1q_scale],
dim=0,
sizes=get_local_sizes(),
)
if self.use_dp:
topk_weights, topk_ids, a1q, a1q_scale = get_dp_group().all_gatherv(
[topk_weights, topk_ids, a1q, a1q_scale],
dim=0,
sizes=get_local_sizes(),
)
if quant_config.quant_dtype == "nvfp4":
a1q_scale = nvfp4_block_scale_interleave(a1q_scale)

8
vllm/model_executor/layers/quantization/modelopt.py
View File

@@ -567,9 +567,13 @@ class ModelOptFp8MoEMethod(FusedMoEMethodBase):
return fp8_w8a8_moe_quant_config(
w1_scale=layer.w13_weight_scale,
g1_alphas=(layer.w13_weight_scale * layer.w13_input_scale).squeeze(),
w2_scale=layer.w2_weight_scale,
g2_alphas=(layer.w2_weight_scale * layer.w2_input_scale).squeeze(),
a1_scale=layer.w13_input_scale,
a1_gscale=layer.w13_input_scale,
a2_scale=layer.w2_input_scale,
a2_gscale=1.0 / layer.w2_input_scale,
per_act_token_quant=False,
)
@@ -1138,8 +1142,8 @@ class ModelOptNvFp4FusedMoE(FusedMoEMethodBase):
moe: FusedMoEConfig,
layer: torch.nn.Module,
) -> None:
from vllm.model_executor.layers.quantization.utils.nvfp4_moe_support import ( # noqa: E501
detect_nvfp4_moe_support,
from vllm.model_executor.layers.quantization.utils.nvfp4_moe_support import (
detect_nvfp4_moe_support, # noqa: E501
)
super().__init__(moe)