support mxfp8 quantization (qwen dense) (#5723)
### What this PR does / why we need it?
support mxfp8 quantization (qwen liner layer)
### Does this PR introduce _any_ user-facing change?
### How was this patch tested?
- vLLM version: v0.13.0
- vLLM main:
2f4e6548ef
Signed-off-by: wangyao <iwangyao@outlook.com>
This commit is contained in:
@@ -45,7 +45,7 @@ from vllm_ascend.ops.linear import AscendUnquantizedLinearMethod
|
|||||||
from vllm_ascend.utils import (ASCEND_QUANTIZATION_METHOD, flashcomm2_enable,
|
from vllm_ascend.utils import (ASCEND_QUANTIZATION_METHOD, flashcomm2_enable,
|
||||||
mlp_tp_enable, oproj_tp_enable)
|
mlp_tp_enable, oproj_tp_enable)
|
||||||
|
|
||||||
from .utils import get_quant_method
|
from .utils import get_quant_method, is_mx_quant_type
|
||||||
|
|
||||||
|
|
||||||
@register_quantization_config(ASCEND_QUANTIZATION_METHOD)
|
@register_quantization_config(ASCEND_QUANTIZATION_METHOD)
|
||||||
@@ -401,7 +401,8 @@ class AscendLinearMethod(LinearMethodBase):
|
|||||||
set_weight_attrs(param, {"output_dim": 0})
|
set_weight_attrs(param, {"output_dim": 0})
|
||||||
layer.register_parameter(pergroup_name, param)
|
layer.register_parameter(pergroup_name, param)
|
||||||
set_weight_attrs(param, extra_weight_attrs)
|
set_weight_attrs(param, extra_weight_attrs)
|
||||||
if "weight_scale_second" in pergroup_name or "weight_offset_second" in pergroup_name:
|
if "weight_scale_second" in pergroup_name or "weight_offset_second" in pergroup_name \
|
||||||
|
or is_mx_quant_type(self.quant_method):
|
||||||
setattr(param, "input_dim", 1)
|
setattr(param, "input_dim", 1)
|
||||||
param.input_dim = 1
|
param.input_dim = 1
|
||||||
|
|
||||||
|
|||||||
@@ -14,6 +14,7 @@ from .w8a8_dynamic import (AscendW8A8DynamicFusedMoEMethod,
|
|||||||
AscendW8A8DynamicLinearMethod)
|
AscendW8A8DynamicLinearMethod)
|
||||||
from .w8a8_pdmix import (AscendW8A8PDMixFusedMoeMethod,
|
from .w8a8_pdmix import (AscendW8A8PDMixFusedMoeMethod,
|
||||||
AscendW8A8PDMixLinearMethod)
|
AscendW8A8PDMixLinearMethod)
|
||||||
|
from .w8a8mxfp8 import AscendW8A8MXFP8DynamicLinearMethod
|
||||||
from .w8a16 import AscendW8A16LinearMethod
|
from .w8a16 import AscendW8A16LinearMethod
|
||||||
|
|
||||||
ASCEND_QUANTIZATION_METHOD_MAP: Dict[str, Dict[str, Type[Any]]] = {
|
ASCEND_QUANTIZATION_METHOD_MAP: Dict[str, Dict[str, Type[Any]]] = {
|
||||||
@@ -40,7 +41,10 @@ ASCEND_QUANTIZATION_METHOD_MAP: Dict[str, Dict[str, Type[Any]]] = {
|
|||||||
},
|
},
|
||||||
"W8A16": {
|
"W8A16": {
|
||||||
"linear": AscendW8A16LinearMethod,
|
"linear": AscendW8A16LinearMethod,
|
||||||
}
|
},
|
||||||
|
"W8A8_MXFP8": {
|
||||||
|
"linear": AscendW8A8MXFP8DynamicLinearMethod,
|
||||||
|
},
|
||||||
}
|
}
|
||||||
|
|
||||||
|
|
||||||
@@ -113,3 +117,9 @@ def get_quant_method_modelslim(
|
|||||||
)
|
)
|
||||||
raise NotImplementedError("Currently, vLLM Ascend only supports following quant types:" \
|
raise NotImplementedError("Currently, vLLM Ascend only supports following quant types:" \
|
||||||
f"{list(ASCEND_QUANTIZATION_METHOD_MAP.keys())}")
|
f"{list(ASCEND_QUANTIZATION_METHOD_MAP.keys())}")
|
||||||
|
|
||||||
|
|
||||||
|
def is_mx_quant_type(instance: Any) -> bool:
|
||||||
|
"""Checks if the quantization method is a mix-precision type."""
|
||||||
|
MX_QUANT_TYPES = (AscendW8A8MXFP8DynamicLinearMethod, )
|
||||||
|
return isinstance(instance, MX_QUANT_TYPES)
|
||||||
|
|||||||
98
vllm_ascend/quantization/w8a8mxfp8.py
Normal file
98
vllm_ascend/quantization/w8a8mxfp8.py
Normal file
@@ -0,0 +1,98 @@
|
|||||||
|
#
|
||||||
|
# Copyright (c) 2025 Huawei Technologies Co., Ltd. All Rights Reserved.
|
||||||
|
# This file is a part of the vllm-ascend project.
|
||||||
|
#
|
||||||
|
# Licensed under the Apache License, Version 2.0 (the "License");
|
||||||
|
# you may not use this file except in compliance with the License.
|
||||||
|
# You may obtain a copy of the License at
|
||||||
|
#
|
||||||
|
# http://www.apache.org/licenses/LICENSE-2.0
|
||||||
|
#
|
||||||
|
# Unless required by applicable law or agreed to in writing, software
|
||||||
|
# distributed under the License is distributed on an "AS IS" BASIS,
|
||||||
|
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
||||||
|
# See the License for the specific language governing permissions and
|
||||||
|
# limitations under the License.
|
||||||
|
#
|
||||||
|
|
||||||
|
from typing import Any, Dict, Optional
|
||||||
|
|
||||||
|
import torch
|
||||||
|
import torch_npu
|
||||||
|
from vllm.config import get_current_vllm_config
|
||||||
|
|
||||||
|
|
||||||
|
class AscendW8A8MXFP8DynamicLinearMethod:
|
||||||
|
"""Linear method for Ascend W8A8_DYNAMIC.
|
||||||
|
"""
|
||||||
|
model_dtype = None
|
||||||
|
|
||||||
|
def __init__(self):
|
||||||
|
vllm_config = get_current_vllm_config()
|
||||||
|
self.group_size = vllm_config.quant_config.quant_description.get(
|
||||||
|
"group_size", 32)
|
||||||
|
|
||||||
|
@staticmethod
|
||||||
|
def get_weight(input_size: int, output_size: int,
|
||||||
|
params_dtype: torch.dtype) -> Dict[str, Any]:
|
||||||
|
params_dict = {
|
||||||
|
"weight":
|
||||||
|
torch.empty(output_size, input_size, dtype=torch.float8_e4m3fn)
|
||||||
|
}
|
||||||
|
return params_dict
|
||||||
|
|
||||||
|
@staticmethod
|
||||||
|
def get_pertensor_param(params_dtype: torch.dtype) -> Dict[str, Any]:
|
||||||
|
return {}
|
||||||
|
|
||||||
|
@staticmethod
|
||||||
|
def get_perchannel_param(
|
||||||
|
output_size: int,
|
||||||
|
params_dtype: torch.dtype,
|
||||||
|
) -> Dict[str, Any]:
|
||||||
|
return {}
|
||||||
|
|
||||||
|
def get_pergroup_param(self,
|
||||||
|
input_size: int,
|
||||||
|
output_size: int,
|
||||||
|
params_dtype: torch.dtype,
|
||||||
|
layer_type: Optional[str] = None) -> Dict[str, Any]:
|
||||||
|
params_dict = {}
|
||||||
|
params_dict["weight_scale"] = torch.empty(output_size,
|
||||||
|
input_size //
|
||||||
|
self.group_size,
|
||||||
|
dtype=torch.uint8)
|
||||||
|
return params_dict
|
||||||
|
|
||||||
|
def apply(
|
||||||
|
self,
|
||||||
|
layer: torch.nn.Module,
|
||||||
|
x: torch.Tensor,
|
||||||
|
bias: Optional[torch.Tensor] = None,
|
||||||
|
tp_rank: Optional[int] = 0,
|
||||||
|
) -> torch.Tensor:
|
||||||
|
|
||||||
|
quantized_x, dynamic_scale = torch_npu.npu_dynamic_mx_quant(
|
||||||
|
x, dst_type=torch.float8_e4m3fn)
|
||||||
|
pertoken_scale = dynamic_scale
|
||||||
|
output_dtype = x.dtype
|
||||||
|
|
||||||
|
output = torch_npu.npu_quant_matmul(
|
||||||
|
quantized_x,
|
||||||
|
layer.weight,
|
||||||
|
layer.weight_scale,
|
||||||
|
scale_dtype=torch_npu.float8_e8m0fnu,
|
||||||
|
pertoken_scale=pertoken_scale,
|
||||||
|
pertoken_scale_dtype=torch_npu.float8_e8m0fnu,
|
||||||
|
bias=bias,
|
||||||
|
output_dtype=output_dtype,
|
||||||
|
group_sizes=[1, 1, self.group_size])
|
||||||
|
|
||||||
|
return output
|
||||||
|
|
||||||
|
def process_weights_after_loading(self, layer):
|
||||||
|
n_dim, k_dim = layer.weight_scale.data.shape
|
||||||
|
layer.weight_scale.data = layer.weight_scale.data.reshape(
|
||||||
|
n_dim, k_dim // 2, 2)
|
||||||
|
layer.weight.data = layer.weight.data.transpose(0, 1)
|
||||||
|
layer.weight_scale.data = layer.weight_scale.data.transpose(0, 1)
|
||||||
Reference in New Issue
Block a user