pytorch默认初始化_Pytorch参数初始化--默认与自定义
1. Pytorch module默认初始化
Pytorch 的Module都只带默认的初始化方法，而且初始化会调用此函数，因此我们定义好后，不用手动初始化。
Conv
class _ConvNd(Module):
__constants__ = ['stride', 'padding', 'dilation', 'groups', 'bias',
'padding_mode', 'output_padding', 'in_channels',
'out_channels', 'kernel_size']
def __init__(self, in_channels, out_channels, kernel_size, stride,
padding, dilation, transposed, output_padding,
groups, bias, padding_mode):
super(_ConvNd, self).__init__()
if in_channels % groups != 0:
raise ValueError('in_channels must be divisible by groups')
if out_channels % groups != 0:
raise ValueError('out_channels must be divisible by groups')
self.in_channels = in_channels
self.out_channels = out_channels
self.kernel_size = kernel_size
self.stride = stride
self.padding = padding
self.dilation = dilation
self.transposed = transposed
self.output_padding = output_padding
self.groups = groups
self.padding_mode = padding_mode
if transposed:
self.weight = Parameter(torch.Tensor(
in_channels, out_channels // groups, *kernel_size))
else:
self.weight = Parameter(torch.Tensor(
out_channels, in_channels // groups, *kernel_size))
if bias:
self.bias = Parameter(torch.Tensor(out_channels))
else:
self.register_parameter('bias', None)
self.reset_parameters()
def reset_parameters(self):
init.kaiming_uniform_(self.weight, a=math.sqrt(5))
if self.bias is not None:
fan_in, _ = init._calculate_fan_in_and_fan_out(self.weight)
bound = 1 / math.sqrt(fan_in)
init.uniform_(self.bias, -bound, bound)
如上图所示，在__init__中最后一行调用函数reset_parameters进行参数初始化，卷积函数都继承了_ConvNd，因此所有卷积module都自动初始化。
我的Pytorch版本是1.2，此版本的初始化函数还是用的何凯名大神的kaiming_uniform_，真的牛逼。
Linear
class Linear(Module):
r"""Applies a linear transformation to the incoming data: :math:`y = xA^T + b`
Args:
in_features: size of each input sample
out_features: size of each output sample
bias: If set to ``False``, the layer will not learn an additive bias.
Default: ``True``
Shape:
- Input: :math:`(N, *, H_{in})` where :math:`*` means any number of
additional dimensions and :math:`H_{in} = \text{in\_features}`
- Output: :math:`(N, *, H_{out})` where all but the last dimension
are the same shape as the input and :math:`H_{out} = \text{out\_features}`.
Attributes:
weight: the learnable weights of the module of shape
:math:`(\text{out\_features}, \text{in\_features})`. The values are
initialized from :math:`\mathcal{U}(-\sqrt{k}, \sqrt{k})`, where
:math:`k = \frac{1}{\text{in\_features}}`
bias: the learnable bias of the module of shape :math:`(\text{out\_features})`.
If :attr:`bias` is ``True``, the values are initialized from
:math:`\mathcal{U}(-\sqrt{k}, \sqrt{k})` where
:math:`k = \frac{1}{\text{in\_features}}`
Examples::
>>> m = nn.Linear(20, 30)
>>> input = torch.randn(128, 20)
>>> output = m(input)
>>> print(output.size())
torch.Size([128, 30])
"""
__constants__ = ['bias', 'in_features', 'out_features']
def __init__(self, in_features, out_features, bias=True):
super(Linear, self).__init__()
self.in_features = in_features
self.out_features = out_features
self.weight = Parameter(torch.Tensor(out_features, in_features))
if bias:
self.bias = Parameter(torch.Tensor(out_features))
else:
self.register_parameter('bias', None)
self.reset_parameters()
def reset_parameters(self):
init.kaiming_uniform_(self.weight, a=math.sqrt(5))
if self.bias is not None:
fan_in, _ = init._calculate_fan_in_and_fan_out(self.weight)
bound = 1 / math.sqrt(fan_in)
init.uniform_(self.bias, -bound, bound)
如上图所示，全连接的初始化方法和卷积也是一样的。
2. 自定义
pytorch中提供了多种初始化函数：
torch.nn.init.constant(tensor, val)
torch.nn.init.normal(tensor, mean=0, std=1)
torch.nn.init.xavier_uniform(tensor, gain=1)
kaiming_uniform_(tensor, a=0, mode='fan_in', nonlinearity='leaky_relu')
可以使用apply()函数初始化，也可以自行对每一个参数初始化：
def weights_init(m):
classname=m.__class__.__name__
if classname.find('Conv') != -1:
xavier(m.weight.data)
xavier(m.bias.data)
net = Net()
net.apply(weights_init) #apply函数会递归地搜索网络内的所有module并把参数表示的函数应用到所有的module上。
def weights_init(m):
if isinstance(m, nn.Conv2d):
xavier(m.weight.data)
xavier(m.bias.data)
Reference