def __init__(self,
input_channels,
output_channels,
kernel_size=-1,
stride=-1,
padding=-1,
groups=1,
channel_shuffle=0,
shuffle_groups=1,
last_relu=0,
abits=8,
wbits=8,
first_layer=0):
super(DorefaConv2d, self).__init__()
self.last_relu = last_relu
self.channel_shuffle_flag = channel_shuffle
self.shuffle_groups = shuffle_groups
self.first_layer = first_layer
self.q_conv = Conv2d_Q(input_channels,
output_channels,
kernel_size=kernel_size,
stride=stride,
padding=padding,
groups=groups,
a_bits=abits,
w_bits=wbits,
first_layer=first_layer)
self.bn = nn.BatchNorm2d(output_channels)
self.relu = nn.ReLU(inplace=True)
def __init__(self, input_channels, output_channels,
kernel_size=-1, stride=-1, padding=-1, groups=1, channel_shuffle=0, shuffle_groups=1, last_relu=0, abits=8, wbits=8, bn_fold=0, q_type=1, first_layer=0):
super(QuanConv2d, self).__init__()
self.last_relu = last_relu
self.channel_shuffle_flag = channel_shuffle
self.shuffle_groups = shuffle_groups
self.bn_fold = bn_fold
self.first_layer = first_layer
if self.bn_fold == 1:
self.bn_q_conv = BNFold_Conv2d_Q(input_channels, output_channels,
kernel_size=kernel_size, stride=stride, padding=padding, groups=groups, a_bits=abits, w_bits=wbits, q_type=q_type, first_layer=first_layer)
else:
self.q_conv = Conv2d_Q(input_channels, output_channels,
kernel_size=kernel_size, stride=stride, padding=padding, groups=groups, a_bits=abits, w_bits=wbits, q_type=q_type, first_layer=first_layer)
self.bn = nn.BatchNorm2d(output_channels, momentum=0.01) # 考虑量化带来的抖动影响,对momentum进行调整(0.1 ——> 0.01),削弱batch统计参数占比，一定程度抑制抖动。经实验量化训练效果更好,acc提升1%左右
self.relu = nn.ReLU(inplace=True)
def __init__(self,
input_channels,
output_channels,
kernel_size=-1,
stride=-1,
padding=-1,
last_relu=0,
groups=1,
channel_shuffle=0,
shuffle_groups=1,
abits=8,
wbits=8,
bn_fold=0,
q_type=1,
first_layer=0):
super(QuanConv2d, self).__init__()
self.last_relu = last_relu
self.channel_shuffle_flag = channel_shuffle
self.shuffle_groups = shuffle_groups
self.bn_fold = bn_fold
self.first_layer = first_layer
if self.bn_fold == 1:
self.bn_q_conv = BNFold_Conv2d_Q(input_channels,
output_channels,
kernel_size=kernel_size,
stride=stride,
padding=padding,
groups=groups,
a_bits=abits,
w_bits=wbits,
q_type=q_type,
first_layer=first_layer)
else:
self.q_conv = Conv2d_Q(input_channels,
output_channels,
kernel_size=kernel_size,
stride=stride,
padding=padding,
groups=groups,
a_bits=abits,
w_bits=wbits,
q_type=q_type,
first_layer=first_layer)
self.bn = nn.BatchNorm2d(output_channels)
self.relu = nn.ReLU(inplace=True)