YOLOV5源码的下载：
git clone https://github.com/ultralytics/yolov5.git
YOLOV5代码yolo.py文件解读:
import argparse
import logging
import sys
from copy import deepcopy
from pathlib import Pathimport mathsys.path.append('./') # to run '$ python *.py' files in subdirectories
logger = logging.getLogger(__name__)import torch
import torch.nn as nnfrom models.common import Conv, Bottleneck, SPP, DWConv, Focus, BottleneckCSP, Concat, NMS, autoShape
from models.experimental import MixConv2d, CrossConv, C3
from utils.general import check_anchor_order, make_divisible, check_file, set_logging
from utils.torch_utils import time_synchronized, fuse_conv_and_bn, model_info, scale_img, initialize_weights, \select_device, copy_attrclass Detect(nn.Module):stride = None # strides computed during buildexport = False # onnx exportdef __init__(self, nc=80, anchors=(), ch=()): # detection layersuper(Detect, self).__init__()self.nc = nc # number of classesself.no = nc + 5 # number of outputs per anchor. VOC: 20+5=25self.nl = len(anchors) # number of detection layers = 3self.na = len(anchors[0]) // 2 # number of anchors =3self.grid = [torch.zeros(1)] * self.nl # init grida = torch.tensor(anchors).float().view(self.nl, -1, 2)# 模型中需要保存下来的参数包括两种: 一种是反向传播需要被optimizer更新的，称之为 parameter;# 一种是反向传播不需要被optimizer更新，称之为 buffer。# 第二种参数我们需要创建tensor, 然后将tensor通过register_buffer()进行注册，# 可以通过model.buffers() 返回，注册完后参数也会自动保存到OrderDict中去。# 注意：buffer的更新在forward中，optim.step只能更新nn.parameter类型的参数self.register_buffer('anchors', a) # shape(nl,na,2)self.register_buffer('anchor_grid', a.clone().view(self.nl, 1, -1, 1, 1, 2)) # shape(nl,1,na,1,1,2)self.m = nn.ModuleList(nn.Conv2d(x, self.no * self.na, 1) for x in ch) # output conv 1*1卷积def forward(self, x):# x = x.copy() # for profilingz = [] # inference outputself.training |= self.exportfor i in range(self.nl):x[i] = self.m[i](x[i]) # convbs, _, ny, nx = x[i].shape # x(bs,255,20,20) to x(bs,3,20,20,85)x[i] = x[i].view(bs, self.na, self.no, ny, nx).permute(0, 1, 3, 4, 2).contiguous()if not self.training: # inferenceif self.grid[i].shape[2:4] != x[i].shape[2:4]:self.grid[i] = self._make_grid(nx, ny).to(x[i].device)y = x[i].sigmoid()y[..., 0:2] = (y[..., 0:2] * 2. - 0.5 + self.grid[i].to(x[i].device)) * self.stride[i] # xyy[..., 2:4] = (y[..., 2:4] * 2) ** 2 * self.anchor_grid[i] # whz.append(y.view(bs, -1, self.no)) # 预测框坐标信息return x if self.training else (torch.cat(z, 1), x) # 预测框坐标, obj, cls@staticmethoddef _make_grid(nx=20, ny=20):# 划分为单元网格yv, xv = torch.meshgrid([torch.arange(ny), torch.arange(nx)])return torch.stack((xv, yv), 2).view((1, 1, ny, nx, 2)).float()# 网络模型类
class Model(nn.Module):def __init__(self, cfg='yolov5s.yaml', ch=3, nc=None): # model, input channels, number of classessuper(Model, self).__init__()if isinstance(cfg, dict):self.yaml = cfg # model dictelse: # is *.yamlimport yaml # for torch hubself.yaml_file = Path(cfg).namewith open(cfg) as f:self.yaml = yaml.load(f, Loader=yaml.FullLoader) # model dict# Define modelif nc and nc != self.yaml['nc']:print('Overriding model.yaml nc=%g with nc=%g' % (self.yaml['nc'], nc))self.yaml['nc'] = nc # override yaml valueself.model, self.save = parse_model(deepcopy(self.yaml), ch=[ch]) # model, savelist, ch_out# print([x.shape for x in self.forward(torch.zeros(1, ch, 64, 64))])# Build strides, anchorsm = self.model[-1] # Detect()if isinstance(m, Detect):s = 128 # 2x min stride# m.stride = [8,16,32]m.stride = torch.tensor([s / x.shape[-2] for x in self.forward(torch.zeros(1, ch, s, s))]) # forward # anchor大小计算, 例如 [10, 13] --> [1.25, 1.625]m.anchors /= m.stride.view(-1, 1, 1)check_anchor_order(m) # 检查anchor顺序和stride顺序是否一致self.stride = m.strideself._initialize_biases() # 初始化偏置 only run once# print('Strides: %s' % m.stride.tolist())# Init weights, biasesinitialize_weights(self) # 初始化权重self.info()print('')def forward(self, x, augment=False, profile=False):if augment: # TTA (Test Time Augmentation)img_size = x.shape[-2:] # height, widths = [1, 0.83, 0.67] # scalesf = [None, 3, None] # flips (2-ud, 3-lr)y = [] # outputsfor si, fi in zip(s, f):xi = scale_img(x.flip(fi) if fi else x, si) # 改变图像尺寸yi = self.forward_once(xi)[0] # forward# cv2.imwrite('img%g.jpg' % s, 255 * xi[0].numpy().transpose((1, 2, 0))[:, :, ::-1]) # saveyi[..., :4] /= si # de-scaleif fi == 2:yi[..., 1] = img_size[0] - yi[..., 1] # de-flip udelif fi == 3:yi[..., 0] = img_size[1] - yi[..., 0] # de-flip lry.append(yi)return torch.cat(y, 1), None # augmented inference, trainelse:return self.forward_once(x, profile) # single-scale inference, traindef forward_once(self, x, profile=False):y, dt = [], [] # outputsfor m in self.model:if m.f != -1: # if not from previous layerx = y[m.f] if isinstance(m.f, int) else [x if j == -1 else y[j] for j in m.f] # from earlier layersif profile:try:import thop # THOP: PyTorch-OpCounter 估算PyTorch模型的FLOPs模块o = thop.profile(m, inputs=(x,), verbose=False)[0] / 1E9 * 2 # FLOPSexcept:o = 0t = time_synchronized()for _ in range(10):_ = m(x)dt.append((time_synchronized() - t) * 100)print('%10.1f%10.0f%10.1fms %-40s' % (o, m.np, dt[-1], m.type))x = m(x) # 执行网络组件操作y.append(x if m.i in self.save else None) # save outputif profile:print('%.1fms total' % sum(dt))return xdef _initialize_biases(self, cf=None): # initialize biases into Detect(), cf is class frequency# https://arxiv.org/abs/1708.02002 section 3.3# cf = torch.bincount(torch.tensor(np.concatenate(dataset.labels, 0)[:, 0]).long(), minlength=nc) + 1.m = self.model[-1] # Detect() modulefor mi, s in zip(m.m, m.stride): # fromb = mi.bias.view(m.na, -1) # conv.bias(255) to (3,85)b[:, 4] += math.log(8 / (640 / s) ** 2) # obj (8 objects per 640 image)b[:, 5:] += math.log(0.6 / (m.nc - 0.99)) if cf is None else torch.log(cf / cf.sum()) # clsmi.bias = torch.nn.Parameter(b.view(-1), requires_grad=True)def _print_biases(self):m = self.model[-1] # Detect() modulefor mi in m.m: # fromb = mi.bias.detach().view(m.na, -1).T # conv.bias(255) to (3,85)print(('%6g Conv2d.bias:' + '%10.3g' * 6) % (mi.weight.shape[1], *b[:5].mean(1).tolist(), b[5:].mean()))# def _print_weights(self):# for m in self.model.modules():# if type(m) is Bottleneck:# print('%10.3g' % (m.w.detach().sigmoid() * 2)) # shortcut weightsdef fuse(self): # fuse model Conv2d() + BatchNorm2d() layersprint('Fusing layers... ')for m in self.model.modules():if type(m) is Conv and hasattr(m, 'bn'):m._non_persistent_buffers_set = set() # pytorch 1.6.0 compatabilitym.conv = fuse_conv_and_bn(m.conv, m.bn) # update convdelattr(m, 'bn') # remove batchnormm.forward = m.fuseforward # update forwardself.info()return selfdef nms(self, mode=True): # add or remove NMS modulepresent = type(self.model[-1]) is NMS # last layer is NMSif mode and not present:print('Adding NMS... ')m = NMS() # modulem.f = -1 # fromm.i = self.model[-1].i + 1 # indexself.model.add_module(name='%s' % m.i, module=m) # addself.eval()elif not mode and present:print('Removing NMS... ')self.model = self.model[:-1] # removereturn selfdef autoshape(self): # add autoShape moduleprint('Adding autoShape... ')m = autoShape(self) # wrap modelcopy_attr(m, self, include=('yaml', 'nc', 'hyp', 'names', 'stride'), exclude=()) # copy attributesreturn mdef info(self, verbose=False): # print model informationmodel_info(self, verbose)# 解析网络模型配置文件并构建模型
def parse_model(d, ch): # model_dict, input_channels(3)logger.info('\n%3s%18s%3s%10s %-40s%-30s' % ('', 'from', 'n', 'params', 'module', 'arguments'))#将模型结构的depth_multiple，width_multiple提取出，赋值给gd (yolov5s: 0.33)，gw (yolov5s:0.50)anchors, nc, gd, gw = d['anchors'], d['nc'], d['depth_multiple'], d['width_multiple'] na = (len(anchors[0]) // 2) if isinstance(anchors, list) else anchors # number of anchors =3no = na * (nc + 5) # number of outputs = anchors * (classes + 5); VOC : 75layers, save, c2 = [], [], ch[-1] # layers, savelist, ch out = 3for i, (f, n, m, args) in enumerate(d['backbone'] + d['head']): # from, number, module, argsm = eval(m) if isinstance(m, str) else m # eval stringsfor j, a in enumerate(args):try:args[j] = eval(a) if isinstance(a, str) else a # eval stringsexcept:pass# 控制深度的代码n = max(round(n * gd), 1) if n > 1 else n # depth gainif m in [Conv, Bottleneck, SPP, DWConv, MixConv2d, Focus, CrossConv, BottleneckCSP, C3]:c1, c2 = ch[f], args[0]# Normal# if i > 0 and args[0] != no: # channel expansion factor# ex = 1.75 # exponential (default 2.0)# e = math.log(c2 / ch[1]) / math.log(2)# c2 = int(ch[1] * ex ** e)# if m != Focus:# 控制宽度(卷积核个数)的代码c2 = make_divisible(c2 * gw, 8) if c2 != no else c2# Experimental# if i > 0 and args[0] != no: # channel expansion factor# ex = 1 + gw # exponential (default 2.0)# ch1 = 32 # ch[1]# e = math.log(c2 / ch1) / math.log(2) # level 1-n# c2 = int(ch1 * ex ** e)# if m != Focus:# c2 = make_divisible(c2, 8) if c2 != no else c2args = [c1, c2, *args[1:]]if m in [BottleneckCSP, C3]:args.insert(2, n)n = 1elif m is nn.BatchNorm2d:args = [ch[f]]elif m is Concat:c2 = sum([ch[-1 if x == -1 else x + 1] for x in f])elif m is Detect:args.append([ch[x + 1] for x in f])if isinstance(args[1], int): # number of anchorsargs[1] = [list(range(args[1] * 2))] * len(f)else:c2 = ch[f]# *args表示接收任意个数量的参数，调用时会将实际参数打包为一个元组传入实参m_ = nn.Sequential(*[m(*args) for _ in range(n)]) if n > 1 else m(*args) # modulet = str(m)[8:-2].replace('__main__.', '') # module typenp = sum([x.numel() for x in m_.parameters()]) # number paramsm_.i, m_.f, m_.type, m_.np = i, f, t, np # attach index, 'from' index, type, number paramslogger.info('%3s%18s%3s%10.0f %-40s%-30s' % (i, f, n, np, t, args)) # printsave.extend(x % i for x in ([f] if isinstance(f, int) else f) if x != -1) # append to savelistlayers.append(m_)ch.append(c2)return nn.Sequential(*layers), sorted(save)if __name__ == '__main__':# 建立参数解析对象parserparser = argparse.ArgumentParser()# 添加属性：给xx实例增加一个aa属性，如 xx.add_argument("aa")parser.add_argument('--cfg', type=str, default='yolov5s.yaml', help='model.yaml')parser.add_argument('--device', default='', help='cuda device, i.e. 0 or 0,1,2,3 or cpu')# 采用parser对象的parse_args函数获取解析的参数opt = parser.parse_args()opt.cfg = check_file(opt.cfg) # check fileset_logging()device = select_device(opt.device)# Create modelmodel = Model(opt.cfg).to(device)model.train()# Profile# img = torch.rand(8 if torch.cuda.is_available() else 1, 3, 640, 640).to(device)# y = model(img, profile=True)# Tensorboard# from torch.utils.tensorboard import SummaryWriter# tb_writer = SummaryWriter()# print("Run 'tensorboard --logdir=models/runs' to view tensorboard at http://localhost:6006/")# tb_writer.add_graph(model.model, img) # add model to tensorboard# tb_writer.add_image('test', img[0], dataformats='CWH') # add model to tensorboard
