Skip to content

[Quantization] Support w4afp8 moe weight offline permute & load #5613

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
Jiang-Jia-Jun merged 2 commits into from
Dec 22, 2025
Merged

[Quantization] Support w4afp8 moe weight offline permute & load #5613

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
01deb9a
support w4afp8 moe offline permute & load
Sunny-bot1 Dec 17, 2025
96c036c
Merge branch 'develop' into w4afp8_offline
Sunny-bot1 Dec 21, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
99 changes: 63 additions & 36 deletions fastdeploy/model_executor/layers/moe/fused_moe_cutlass_backend.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -772,8 +772,9 @@ def process_prequanted_weights(self, layer: nn.Layer, state_dict, is_rearrange:
down_proj_expert_weight_key = layer.weight_key_map.get("down_proj_expert_weight_key", None)
up_gate_proj_expert_weight_scale_key = layer.weight_key_map.get("up_gate_proj_expert_weight_scale_key", None)
down_proj_expert_weight_scale_key = layer.weight_key_map.get("down_proj_expert_weight_scale_key", None)
up_gate_proj_expert_in_scale_key = layer.weight_key_map.get("up_gate_proj_expert_in_scale_key", None)
down_proj_expert_in_scale_key = layer.weight_key_map.get("down_proj_expert_in_scale_key", None)
if not layer.moe_quant_config.moe_dynamic_quant:
up_gate_proj_expert_in_scale_key = layer.weight_key_map.get("up_gate_proj_expert_in_scale_key", None)
down_proj_expert_in_scale_key = layer.weight_key_map.get("down_proj_expert_in_scale_key", None)

up_gate_proj_weights, down_proj_weights, logical_expert_ids, ep_rank_to_expert_id_list = (
layer.load_experts_weight(
Expand All @@ -793,7 +794,7 @@ def process_prequanted_weights(self, layer: nn.Layer, state_dict, is_rearrange:
if isinstance(state_dict, list):
state_dict = dict(state_dict)

if layer.ep_size > 1:
if layer.ep_size > 1 and not layer.moe_quant_config.moe_dynamic_quant:
for expert_idx in ep_rank_to_expert_id_list:
scale_tensor = get_tensor(
(
Expand Down Expand Up @@ -826,44 +827,54 @@ def process_prequanted_weights(self, layer: nn.Layer, state_dict, is_rearrange:
layer.fd_config.model_config.model,
)
)
up_gate_proj_in_scale.append(
get_tensor(
(
state_dict.pop(up_gate_proj_expert_in_scale_key.format(expert_idx))
if up_gate_proj_expert_in_scale_key.format(expert_idx) in state_dict
else up_gate_proj_expert_in_scale_key.format(expert_idx)
),
layer.fd_config.model_config.model,
if not layer.moe_quant_config.moe_dynamic_quant:
up_gate_proj_in_scale.append(
get_tensor(
(
state_dict.pop(up_gate_proj_expert_in_scale_key.format(expert_idx))
if up_gate_proj_expert_in_scale_key.format(expert_idx) in state_dict
else up_gate_proj_expert_in_scale_key.format(expert_idx)
),
layer.fd_config.model_config.model,
)
)
)
down_proj_in_scale.append(
get_tensor(
(
state_dict.pop(down_proj_expert_in_scale_key.format(expert_idx))
if down_proj_expert_in_scale_key.format(expert_idx) in state_dict
else down_proj_expert_in_scale_key.format(expert_idx)
),
layer.fd_config.model_config.model,
down_proj_in_scale.append(
get_tensor(
(
state_dict.pop(down_proj_expert_in_scale_key.format(expert_idx))
if down_proj_expert_in_scale_key.format(expert_idx) in state_dict
else down_proj_expert_in_scale_key.format(expert_idx)
),
layer.fd_config.model_config.model,
)
)
)

up_gate_proj_weight = paddle.stack(up_gate_proj_weights, axis=0)
down_proj_weight = paddle.stack(down_proj_weights, axis=0)
up_gate_proj_weight_scale = paddle.stack(up_gate_proj_weight_scale, axis=0)
down_proj_weight_scale = paddle.stack(down_proj_weight_scale, axis=0)
up_gate_proj_in_scale_all_experts = paddle.stack(up_gate_proj_in_scale_all_experts, axis=0).squeeze()
up_gate_proj_in_scale = paddle.stack(up_gate_proj_in_scale, axis=0).squeeze()
down_proj_in_scale = paddle.stack(down_proj_in_scale, axis=0).squeeze()

name_tensor_map = {
"up_gate_proj_weight": up_gate_proj_weight,
"down_proj_weight": down_proj_weight,
"up_gate_proj_weight_scale": up_gate_proj_weight_scale,
"down_proj_weight_scale": down_proj_weight_scale,
"up_gate_proj_in_scale_all_experts": up_gate_proj_in_scale_all_experts,
"up_gate_proj_in_scale": up_gate_proj_in_scale,
"down_proj_in_scale": down_proj_in_scale,
}
if not layer.moe_quant_config.moe_dynamic_quant:
up_gate_proj_in_scale_all_experts = paddle.stack(up_gate_proj_in_scale_all_experts, axis=0).squeeze()
up_gate_proj_in_scale = paddle.stack(up_gate_proj_in_scale, axis=0).squeeze()
down_proj_in_scale = paddle.stack(down_proj_in_scale, axis=0).squeeze()

if not layer.moe_quant_config.moe_dynamic_quant:
name_tensor_map = {
"up_gate_proj_weight": up_gate_proj_weight,
"down_proj_weight": down_proj_weight,
"up_gate_proj_weight_scale": up_gate_proj_weight_scale,
"down_proj_weight_scale": down_proj_weight_scale,
"up_gate_proj_in_scale_all_experts": up_gate_proj_in_scale_all_experts,
"up_gate_proj_in_scale": up_gate_proj_in_scale,
"down_proj_in_scale": down_proj_in_scale,
}
else:
name_tensor_map = {
"up_gate_proj_weight": up_gate_proj_weight,
"down_proj_weight": down_proj_weight,
"up_gate_proj_weight_scale": up_gate_proj_weight_scale,
"down_proj_weight_scale": down_proj_weight_scale,
}
for name, tensor in name_tensor_map.items():
getattr(layer, name).set_value(tensor)

Expand Down Expand Up @@ -1020,11 +1031,27 @@ def create_w4afp8_scale_weights(self, layer: nn.Layer, weight_key_map: dict):

# weight_scales
if layer.is_quantized:
if not layer.moe_quant_config.moe_dynamic_quant:
up_gate_proj_weight_scale_shape = [layer.num_local_experts, layer.moe_intermediate_size * 2]
down_proj_weight_scale_shape = [layer.num_local_experts, layer.hidden_size]
else:
up_gate_proj_weight_scale_shape = [
layer.num_local_experts,
layer.moe_intermediate_size * 2 // 128,
layer.hidden_size // 128,
128,
]
down_proj_weight_scale_shape = [
layer.num_local_experts,
layer.hidden_size // 128,
layer.moe_intermediate_size // 128,
128,
]
setattr(
layer,
"up_gate_proj_weight_scale",
layer.create_parameter(
shape=[layer.num_local_experts, layer.moe_intermediate_size * 2],
shape=up_gate_proj_weight_scale_shape,
dtype="float32",
default_initializer=paddle.nn.initializer.Constant(0),
),
Expand All @@ -1033,7 +1060,7 @@ def create_w4afp8_scale_weights(self, layer: nn.Layer, weight_key_map: dict):
layer,
"down_proj_weight_scale",
layer.create_parameter(
shape=[layer.num_local_experts, layer.hidden_size],
shape=down_proj_weight_scale_shape,
dtype="float32",
default_initializer=paddle.nn.initializer.Constant(0),
),
Expand Down
6 changes: 4 additions & 2 deletions fastdeploy/model_executor/layers/moe/moe.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -212,8 +212,10 @@ def __init__(
self._dtype = self._helper.get_default_dtype()
self.weight_dtype = self._dtype

self.is_quantized = fd_config.model_config.is_quantized and not (
fd_config.quant_config.name() == "mix_quant" and fd_config.quant_config.moe_quant_type is None
self.is_moe_quantized = getattr(self.fd_config.model_config, "is_moe_quantized", False)
self.is_quantized = self.is_moe_quantized or (
fd_config.model_config.is_quantized
and not (fd_config.quant_config.name() == "mix_quant" and fd_config.quant_config.moe_quant_type is None)
)
moe_quant_config = fd_config.quant_config
self.moe_quant_config = moe_quant_config
Expand Down
5 changes: 4 additions & 1 deletion fastdeploy/model_executor/layers/quantization/mix_quant.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -40,6 +40,7 @@ def __init__(
is_quantized: bool = False,
hadamard_block_size: int = 128,
moe_dynamic_quant: bool = False,
is_moe_quantized: bool = False,
) -> None:
super().__init__()
self.dense_quant_type = dense_quant_type
Expand All @@ -59,6 +60,7 @@ def __init__(
self.is_quantized = is_quantized
self.hadamard_block_size = hadamard_block_size
self.moe_dynamic_quant = moe_dynamic_quant
self.is_moe_quantized = is_moe_quantized

def name(self) -> str:
return "mix_quant"
Expand All @@ -76,6 +78,7 @@ def from_config(cls, config: dict) -> "MixQuantConfig":
config.get("is_quantized", False),
config.get("hadamard_block_size", 128),
config.get("moe_dynamic_quant", False),
config.get("is_moe_quantized", False),
)

def get_quant_method(self, layer) -> Optional[QuantMethodBase]:
Expand All @@ -102,7 +105,7 @@ def get_quant_method(self, layer) -> Optional[QuantMethodBase]:
.from_config(
{
"is_permuted": self.is_permuted,
"is_quantized": not self.is_checkpoint_bf16,
"is_quantized": not self.is_checkpoint_bf16 or self.is_moe_quantized,
"hadamard_block_size": self.hadamard_block_size,
}
)
Expand Down
Loading