网络输出是48,
这个效果不好,不知道为啥
growth_rate 是通道增长步长
代码:
import time
from collections import OrderedDict
import torch
import torch.nn as nn
import torch.nn.functional as F
class conv_bn_relu(nn.Module):
def __init__(self, nin, nout, kernel_size, stride, padding, bias=False):
super(conv_bn_relu, self).__init__()
self.conv = nn.Conv2d(nin, nout, kernel_size=kernel_size, stride=stride, padding=padding, bias=bias)
self.batch_norm = nn.BatchNorm2d(nout)
self.relu = nn.ReLU6(True)
# self.relu = mish.Mish()
def forward(self, x):
out = self.conv(x)
out = self.batch_norm(out)
out = self.relu(out)
return out
class Transition_layer(nn.Sequential):
本文讨论了PeleeNet V2网络结构在实际应用中输出为48通道时,表现出的效果不尽如人意。作者指出可能的原因并提及了growth_rate参数,暗示其可能影响网络性能。代码细节未给出。
订阅专栏 解锁全文

被折叠的 条评论
为什么被折叠?



