yolov3网络结构图代码分析

 

import torch
from collections import OrderedDict
import math
import torch.nn as nn
import numpy as np
from tensorboardX import SummaryWriter
def conv2d(filter_in,filter_out,kernel_size):
    pad=(kernel_size-1)//2 if kernel_size else 0
    return nn.Sequential(
        nn.Conv2d(filter_in,filter_out,kernel_size=kernel_size,stride=1,padding=pad,bias=False),
        nn.BatchNorm2d(filter_out),
        nn.LeakyReLU(0.1)
    )

def make_last_layers(filter_list,in_filters,out_filter):
    m=nn.ModuleList([
        conv2d(in_filters,filter_list[0],1),
        conv2d(filter_list[0],filter_list[1],3),
        conv2d(filter_list[1], filter_list[0], 1),
        conv2d(filter_list[0], filter_list[1], 3),
        conv2d(filter_list[1], filter_list[0], 1),
        conv2d(filter_list[0], filter_list[1], 3),
        nn.Conv2d(filter_list[1], out_filter, kernel_size=1,stride=1,padding=0,bias=True)
    ])
    return m
# ModuleList(
#   (0): Sequential(
#     (0): Conv2d(3, 32, kernel_size=(1, 1), stride=(1, 1), bias=False)
#     (1): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
#     (2): LeakyReLU(negative_slope=0.1)
#   )
#   (1): Sequential(
#     (0): Conv2d(32, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
#     (1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
#     (2): LeakyReLU(negative_slope=0.1)
#   )
#   (2): Sequential(
#     (0): Conv2d(64, 32, kernel_size=(1, 1), stride=(1, 1), bias=False)
#     (1): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
#     (2): LeakyReLU(negative_slope=0.1)
#   )
#   (3): Sequential(
#     (0): Conv2d(32, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
#