网站支付怎么做,怎么做徐州市环保局网站备案表,架设网站的目的,如何做产品网站的推广DeeplabV3算法 参数配置关于数据集的配置训练集参数 数据预处理模块DataSet构建模块测试一下数据集去正则化模型加载模块DeepLABV3 参数配置
关于数据集的配置 parser argparse.ArgumentParser()# Datset Optionsparser.add_argument(--data_root, typestr, defa… DeeplabV3算法 参数配置关于数据集的配置训练集参数 数据预处理模块DataSet构建模块测试一下数据集去正则化模型加载模块DeepLABV3 参数配置
关于数据集的配置 parser argparse.ArgumentParser()# Datset Optionsparser.add_argument(--data_root, typestr, defaultrD:/,helppath to Dataset)parser.add_argument(--dataset, typestr, defaultvoc,choices[voc, cityscapes], helpName of dataset)parser.add_argument(--num_classes, typeint, defaultNone,helpnum classes (default: None))# Deeplab Options# 选择模型的架构特征提取模块分为moiblienet或者resnet50parser.add_argument(--model, typestr, defaultdeeplabv3plus_resnet50,choices[deeplabv3_resnet50, deeplabv3plus_resnet50,deeplabv3_resnet101, deeplabv3plus_resnet101,deeplabv3_mobilenet, deeplabv3plus_mobilenet], helpmodel name)parser.add_argument(--separable_conv, actionstore_true, defaultFalse,helpapply separable conv to decoder and aspp)parser.add_argument(--output_stride, typeint, default16, choices[8, 16])训练集参数 # Train Options# 制作测试parser.add_argument(--test_only, actionstore_true, defaultFalse)parser.add_argument(--save_val_results, actionstore_true, defaultFalse,helpsave segmentation results to \./results\)parser.add_argument(--total_itrs, typeint, default60e3,helpepoch number (default: 30k))# 学习率parser.add_argument(--lr, typefloat, default0.01,helplearning rate (default: 0.01))parser.add_argument(--lr_policy, typestr, defaultpoly, choices[poly, step],helplearning rate scheduler policy)parser.add_argument(--step_size, typeint, default10000)parser.add_argument(--crop_val, actionstore_true, defaultFalse,helpcrop validation (default: False))parser.add_argument(--batch_size, typeint, default8,helpbatch size (default: 16))parser.add_argument(--val_batch_size, typeint, default4,helpbatch size for validation (default: 4))parser.add_argument(--crop_size, typeint, default513)# 预训练权重路径parser.add_argument(--ckpt, default./checkpoint/best_deeplabv3_resnet50_voc_os16.pth, typestr,helprestore from checkpoint)parser.add_argument(--continue_training, actionstore_true, defaultTrue)parser.add_argument(--loss_type, typestr, defaultcross_entropy,choices[cross_entropy, focal_loss], helploss type (default: False))parser.add_argument(--gpu_id, typestr, default0,helpGPU ID)# 正则化参数parser.add_argument(--weight_decay, typefloat, default1e-4,helpweight decay (default: 1e-4))parser.add_argument(--random_seed, typeint, default1,helprandom seed (default: 1))parser.add_argument(--print_interval, typeint, default10,helpprint interval of loss (default: 10))parser.add_argument(--val_interval, typeint, default100,helpepoch interval for eval (default: 100))parser.add_argument(--download, actionstore_true, defaultFalse,helpdownload datasets)数据预处理模块
分别针对训练集、验证集、测试集做三种数据增强变换
def get_dataset(opts): Dataset And Augmentationif opts.dataset voc:train_transform et.ExtCompose([#et.ExtResize(sizeopts.crop_size),et.ExtRandomScale((0.5, 2.0)),et.ExtRandomCrop(size(opts.crop_size, opts.crop_size), pad_if_neededTrue),et.ExtRandomHorizontalFlip(),et.ExtToTensor(),et.ExtNormalize(mean[0.485, 0.456, 0.406],std[0.229, 0.224, 0.225]),])if opts.crop_val:val_transform et.ExtCompose([et.ExtResize(opts.crop_size),et.ExtCenterCrop(opts.crop_size),et.ExtToTensor(),et.ExtNormalize(mean[0.485, 0.456, 0.406],std[0.229, 0.224, 0.225]),])else:val_transform et.ExtCompose([et.ExtToTensor(),et.ExtNormalize(mean[0.485, 0.456, 0.406],std[0.229, 0.224, 0.225]),])train_dst VOCSegmentation(rootopts.data_root, yearopts.year,image_settrain, downloadopts.download, transformtrain_transform)val_dst VOCSegmentation(rootopts.data_root, yearopts.year,image_setval, downloadFalse, transformval_transform)if opts.dataset cityscapes:train_transform et.ExtCompose([#et.ExtResize( 512 ),et.ExtRandomCrop(size(opts.crop_size, opts.crop_size)),et.ExtColorJitter( brightness0.5, contrast0.5, saturation0.5 ),et.ExtRandomHorizontalFlip(),et.ExtToTensor(),et.ExtNormalize(mean[0.485, 0.456, 0.406],std[0.229, 0.224, 0.225]),])val_transform et.ExtCompose([#et.ExtResize( 512 ),et.ExtToTensor(),et.ExtNormalize(mean[0.485, 0.456, 0.406],std[0.229, 0.224, 0.225]),])train_dst Cityscapes(rootopts.data_root,splittrain, transformtrain_transform)val_dst Cityscapes(rootopts.data_root,splitval, transformval_transform)return train_dst, val_dstDataSet构建模块 def voc_cmap(N256, normalizedFalse):def bitget(byteval, idx):return ((byteval (1 idx)) ! 0)dtype float32 if normalized else uint8cmap np.zeros((N, 3), dtypedtype)for i in range(N):r g b 0c ifor j in range(8):r r | (bitget(c, 0) 7-j)g g | (bitget(c, 1) 7-j)b b | (bitget(c, 2) 7-j)c c 3cmap[i] np.array([r, g, b])cmap cmap/255 if normalized else cmapreturn cmap
class VOCSegmentation(data.Dataset):Pascal VOC http://host.robots.ox.ac.uk/pascal/VOC/_ Segmentation Dataset.Args:root (string): Root directory of the VOC Dataset.year (string, optional): The dataset year, supports years 2007 to 2012.image_set (string, optional): Select the image_set to use, train, trainval or valdownload (bool, optional): If true, downloads the dataset from the internet andputs it in root directory. If dataset is already downloaded, it is notdownloaded again.transform (callable, optional): A function/transform that takes in an PIL imageand returns a transformed version. E.g, transforms.RandomCropcmap voc_cmap()def __init__(self,root,year2012,image_settrain,downloadFalse,transformNone):is_augFalseif year2012_aug:is_aug Trueyear 2012self.root os.path.expanduser(root)self.year yearself.url DATASET_YEAR_DICT[year][url]self.filename DATASET_YEAR_DICT[year][filename]self.md5 DATASET_YEAR_DICT[year][md5]self.transform transformself.image_set image_setbase_dir DATASET_YEAR_DICT[year][base_dir]voc_root os.path.join(self.root, base_dir)image_dir os.path.join(voc_root, JPEGImages)if download:download_extract(self.url, self.root, self.filename, self.md5)if not os.path.isdir(voc_root):raise RuntimeError(Dataset not found or corrupted. You can use downloadTrue to download it)if is_aug and image_settrain:mask_dir os.path.join(voc_root, SegmentationClassAug)assert os.path.exists(mask_dir), SegmentationClassAug not found, please refer to README.md and prepare it manuallysplit_f os.path.join( self.root, train_aug.txt)#./datasets/data/train_aug.txtelse:mask_dir os.path.join(voc_root, SegmentationClass)splits_dir os.path.join(voc_root, ImageSets/Segmentation)split_f os.path.join(splits_dir, image_set.rstrip(\n) .txt)if not os.path.exists(split_f):raise ValueError(Wrong image_set entered! Please use image_settrain or image_settrainval or image_setval)with open(os.path.join(split_f), r) as f:file_names [x.strip() for x in f.readlines()]self.images [os.path.join(image_dir, x .jpg) for x in file_names]self.masks [os.path.join(mask_dir, x .png) for x in file_names]assert (len(self.images) len(self.masks))def __getitem__(self, index):Args:index (int): IndexReturns:tuple: (image, target) where target is the image segmentation.img Image.open(self.images[index]).convert(RGB)target Image.open(self.masks[index])if self.transform is not None:img, target self.transform(img, target)return img, targetdef __len__(self):return len(self.images)classmethoddef decode_target(cls, mask):decode semantic mask to RGB imagereturn cls.cmap[mask]
测试一下数据集
import numpy as npfrom datasets import VOCSegmentation
from utils import ext_transforms as et
import cv2
train_transform et.ExtCompose([# et.ExtResize(sizeopts.crop_size),et.ExtRandomScale((0.5, 2.0)),et.ExtRandomCrop(size(224, 224), pad_if_neededTrue),et.ExtRandomHorizontalFlip(),et.ExtToTensor(),et.ExtNormalize(mean[0.485, 0.456, 0.406],std[0.229, 0.224, 0.225]),
])
data VOCSegmentation(rootrD:/, year2012, image_settrain, downloadFalse, transformtrain_transform)if __name__ __main__:print(data[0][0].shape)print(data[0][1].shape)res data.decode_target(data[0][1])cv2.imshow(Res,np.array(res))cv2.waitKey(0)去正则化
class Denormalize(object):def __init__(self, mean, std):mean np.array(mean)std np.array(std)self._mean -mean/stdself._std 1/stddef __call__(self, tensor):if isinstance(tensor, np.ndarray):return (tensor - self._mean.reshape(-1,1,1)) / self._std.reshape(-1,1,1)return normalize(tensor, self._mean, self._std)
模型加载模块 # Set up modelmodel_map {deeplabv3_resnet50: network.deeplabv3_resnet50,deeplabv3plus_resnet50: network.deeplabv3plus_resnet50,deeplabv3_resnet101: network.deeplabv3_resnet101,deeplabv3plus_resnet101: network.deeplabv3plus_resnet101,deeplabv3_mobilenet: network.deeplabv3_mobilenet,deeplabv3plus_mobilenet: network.deeplabv3plus_mobilenet}model model_map[opts.model](num_classesopts.num_classes, output_strideopts.output_stride)def deeplabv3_resnet50(num_classes21, output_stride8, pretrained_backboneTrue):Constructs a DeepLabV3 model with a ResNet-50 backbone.Args:num_classes (int): number of classes.output_stride (int): output stride for deeplab.pretrained_backbone (bool): If True, use the pretrained backbone.return _load_model(deeplabv3, resnet50, num_classes, output_strideoutput_stride, pretrained_backbonepretrained_backbone)def deeplabv3_resnet101(num_classes21, output_stride8, pretrained_backboneTrue):Constructs a DeepLabV3 model with a ResNet-101 backbone.Args:num_classes (int): number of classes.output_stride (int): output stride for deeplab.pretrained_backbone (bool): If True, use the pretrained backbone.return _load_model(deeplabv3, resnet101, num_classes, output_strideoutput_stride, pretrained_backbonepretrained_backbone)def deeplabv3_mobilenet(num_classes21, output_stride8, pretrained_backboneTrue, **kwargs):Constructs a DeepLabV3 model with a MobileNetv2 backbone.Args:num_classes (int): number of classes.output_stride (int): output stride for deeplab.pretrained_backbone (bool): If True, use the pretrained backbone.return _load_model(deeplabv3, mobilenetv2, num_classes, output_strideoutput_stride, pretrained_backbonepretrained_backbone)# Deeplab v3def deeplabv3plus_resnet50(num_classes21, output_stride8, pretrained_backboneTrue):Constructs a DeepLabV3 model with a ResNet-50 backbone.Args:num_classes (int): number of classes.output_stride (int): output stride for deeplab.pretrained_backbone (bool): If True, use the pretrained backbone.return _load_model(deeplabv3plus, resnet50, num_classes, output_strideoutput_stride, pretrained_backbonepretrained_backbone)加载模块
def _load_model(arch_type, backbone, num_classes, output_stride, pretrained_backbone):if backbonemobilenetv2:model _segm_mobilenet(arch_type, backbone, num_classes, output_strideoutput_stride, pretrained_backbonepretrained_backbone)elif backbone.startswith(resnet):model _segm_resnet(arch_type, backbone, num_classes, output_strideoutput_stride, pretrained_backbonepretrained_backbone)else:raise NotImplementedErrorreturn modeldef _segm_resnet(name, backbone_name, num_classes, output_stride, pretrained_backbone):if output_stride8:replace_stride_with_dilation[False, True, True]aspp_dilate [12, 24, 36]else:replace_stride_with_dilation[False, False, True]aspp_dilate [6, 12, 18]backbone resnet.__dict__[backbone_name](pretrainedpretrained_backbone,replace_stride_with_dilationreplace_stride_with_dilation)inplanes 2048low_level_planes 256if namedeeplabv3plus:return_layers {layer4: out, layer1: low_level}#classifier DeepLabHeadV3Plus(inplanes, low_level_planes, num_classes, aspp_dilate)elif namedeeplabv3:return_layers {layer4: out}classifier DeepLabHead(inplanes , num_classes, aspp_dilate)#提取网络的第几层输出结果并给一个别名backbone IntermediateLayerGetter(backbone, return_layersreturn_layers)model DeepLabV3(backbone, classifier)return model
DeepLABV3
class DeepLabHeadV3Plus(nn.Module):def __init__(self, in_channels, low_level_channels, num_classes, aspp_dilate[12, 24, 36]):super(DeepLabHeadV3Plus, self).__init__()self.project nn.Sequential( nn.Conv2d(low_level_channels, 48, 1, biasFalse),nn.BatchNorm2d(48),nn.ReLU(inplaceTrue),)self.aspp ASPP(in_channels, aspp_dilate)self.classifier nn.Sequential(nn.Conv2d(304, 256, 3, padding(1,1), biasFalse),nn.BatchNorm2d(256),nn.ReLU(inplaceTrue),nn.Conv2d(256, num_classes, 1))self._init_weight()def forward(self, feature):#print(feature.shape)low_level_feature self.project( feature[low_level] )#return_layers {layer4: out, layer1: low_level}#print(low_level_feature.shape)output_feature self.aspp(feature[out])#print(output_feature.shape)output_feature F.interpolate(output_feature, sizelow_level_feature.shape[2:], modebilinear, align_cornersFalse)#print(output_feature.shape)return self.classifier( torch.cat( [ low_level_feature, output_feature ], dim1 ) )def _init_weight(self):for m in self.modules():if isinstance(m, nn.Conv2d):nn.init.kaiming_normal_(m.weight)elif isinstance(m, (nn.BatchNorm2d, nn.GroupNorm)):nn.init.constant_(m.weight, 1)nn.init.constant_(m.bias, 0)其中空洞融合ASPP模块
class ASPP(nn.Module):def __init__(self, in_channels, atrous_rates):super(ASPP, self).__init__()out_channels 256modules []modules.append(nn.Sequential(nn.Conv2d(in_channels, out_channels, 1, biasFalse),nn.BatchNorm2d(out_channels),nn.ReLU(inplaceTrue)))rate1, rate2, rate3 tuple(atrous_rates)modules.append(ASPPConv(in_channels, out_channels, rate1))modules.append(ASPPConv(in_channels, out_channels, rate2))modules.append(ASPPConv(in_channels, out_channels, rate3))modules.append(ASPPPooling(in_channels, out_channels))self.convs nn.ModuleList(modules)self.project nn.Sequential(nn.Conv2d(5 * out_channels, out_channels, 1, biasFalse),nn.BatchNorm2d(out_channels),nn.ReLU(inplaceTrue),nn.Dropout(0.1),)def forward(self, x):res []for conv in self.convs:#print(conv(x).shape)res.append(conv(x))res torch.cat(res, dim1)return self.project(res)
卷积转深度可分离卷积
def convert_to_separable_conv(module):new_module moduleif isinstance(module, nn.Conv2d) and module.kernel_size[0]1:new_module AtrousSeparableConvolution(module.in_channels,module.out_channels, module.kernel_size,module.stride,module.padding,module.dilation,module.bias)for name, child in module.named_children():new_module.add_module(name, convert_to_separable_conv(child))return new_module
class AtrousSeparableConvolution(nn.Module): Atrous Separable Convolutiondef __init__(self, in_channels, out_channels, kernel_size,stride1, padding0, dilation1, biasTrue):super(AtrousSeparableConvolution, self).__init__()self.body nn.Sequential(# Separable Convnn.Conv2d( in_channels, in_channels, kernel_sizekernel_size, stridestride, paddingpadding, dilationdilation, biasbias, groupsin_channels ),# PointWise Convnn.Conv2d( in_channels, out_channels, kernel_size1, stride1, padding0, biasbias),)self._init_weight()def forward(self, x):return self.body(x)def _init_weight(self):for m in self.modules():if isinstance(m, nn.Conv2d):nn.init.kaiming_normal_(m.weight)elif isinstance(m, (nn.BatchNorm2d, nn.GroupNorm)):nn.init.constant_(m.weight, 1)nn.init.constant_(m.bias, 0)参考代码 -------- 链接: https://pan.baidu.com/s/1E6CD5NQjbKeITacH9VjxAw?pwdjqgk 提取码: jqgk –来自百度网盘超级会员v6的分享
