二、pix2pixHD代码解析(options设置)

pix2pixHD代码解析

一、pix2pixHD代码解析(train.py + test.py)
二、pix2pixHD代码解析(options设置)
三、pix2pixHD代码解析(dataset处理)
四、pix2pixHD代码解析(models搭建)

二、pix2pixHD代码解析(options设置)

base_options.py

import argparse
import os
from util import util
import torch

class BaseOptions():
    def __init__(self):
        self.parser = argparse.ArgumentParser()
        self.initialized = False

    def initialize(self):    
        # experiment specifics
        self.parser.add_argument('--name', type=str, default='label2city', help='name of the experiment. It decides where to store samples and models')  # 实验名称, 它决定在哪里存储样本和模型
        self.parser.add_argument('--gpu_ids', type=str, default='0, 1, 2', help='gpu ids: e.g. 0  0,1,2, 0,2. use -1 for CPU')                           # 要使用的GPU id:0, 1, 2三块GPU
        self.parser.add_argument('--checkpoints_dir', type=str, default='./checkpoints', help='models are saved here')                                   # 检查点路径
        self.parser.add_argument('--model', type=str, default='pix2pixHD', help='which model to use')                                                    # 选择的模型，此处由两个模型可供选择                                         # 选择模型
        self.parser.add_argument('--norm', type=str, default='instance', help='instance normalization or batch normalization')        
        self.parser.add_argument('--use_dropout', action='store_true', help='use dropout for the generator')
        self.parser.add_argument('--data_type', default=32, type=int, choices=[8, 16, 32], help="Supported data type i.e. 8, 16, 32 bit")                # 支持的数据类型，即8、16、32位
        self.parser.add_argument('--verbose', action='store_true', default=False, help='toggles verbose')               # 默认为false,表示之前并无模型保存
        self.parser.add_argument('--fp16', action='store_true', default=False, help='train with AMP')                   # fp16和amp适用于混合精度加速（英伟达自带的apex库）
        self.parser.add_argument('--local_rank', type=int, default=0, help='local rank for distributed training')

        # input/output sizes
        self.parser.add_argument('--batchSize', type=int, default=3, help='input batch size')
        self.parser.add_argument('--loadSize', type=int, default=512, help='scale images to this size')                 # 将图像缩放到这个大小：此处仅设置宽度，高度按照宽高比计算
        self.parser.add_argument('--fineSize', type=int, default=512, help='then crop to this size')                    # 然后裁剪成这个大小
        self.parser.add_argument('--label_nc', type=int, default=0, help='# of input label channels')                   # 标签图通道数
        self.parser.add_argument('--input_nc', type=int, default=3, help='# of input image channels')                   # 输入真实图通道数
        self.parser.add_argument('--output_nc', type=int, default=3, help='# of output image channels')                 # 生成的输出图通道数

        # for setting inputs
        self.parser.add_argument('--dataroot', type=str, default='./datasets/geometry/')                                # geometry或cityscapes
        self.parser.add_argument('--resize_or_crop', type=str, default='scale_width_and_crop', help='scaling and cropping of images at load time [resize_and_crop|crop|scale_width|scale_width_and_crop]')  # 这个要设置好
        self.parser.add_argument('--serial_batches', action='store_true', help='if true, takes images in order to make batches, otherwise takes them randomly')  # 如果为真，则按批次加载，否则随机加载
        self.parser.add_argument('--no_flip', action='store_true', help='if specified, do not flip the images for data argumentation')  # 如果指定，不要为了数据论证而翻转图像。
        self.parser.add_argument('--nThreads', default=2, type=int, help='# threads for loading data')                  # 加载数据的线程
        self.parser.add_argument('--max_dataset_size', type=int, default=float("inf"), help='Maximum number of samples allowed per dataset. If the dataset directory contains more than max_dataset_size, only a subset is loaded.')
                                                                                                                        # 每个数据集允许的最大样本数。如果数据集目录包含超过max_dataset_size，则只加载一个子集

        # for displays
        self.parser.add_argument('--display_winsize', type=int, default=512,  help='display window size')
        self.parser.add_argument('--tf_log', action='store_true', help='if specified, use tensorboard logging. Requires tensorflow installed')

        # for generator
        self.parser.add_argument('--netG', type=str, default='global', help='selects model to use for netG')            # 选择netG使用的模型，默认为global
        self.parser.add_argument('--ngf', type=int, default=64, help='# of gen filters in first conv layer')            # 生成器器在第一conv层的卷积数
        self.parser.add_argument('--n_downsample_global', type=int, default=4, help='number of downsampling layers in netG')  # netG中的下采样层数
        self.parser.add_argument('--n_blocks_global', type=int, default=9, help='number of residual blocks in the global generator network')  # 全局生成器网络中残差块的数量
        self.parser.add_argument('--n_blocks_local', type=int, default=3, help='number of residual blocks in the local enhancer network')
        self.parser.add_argument('--n_local_enhancers', type=int, default=1, help='number of local enhancers to use')   # 要使用的局部增强子的数量
        self.parser.add_argument('--niter_fix_global', type=int, default=0, help='number of epochs that we only train the outmost local enhancer')        

        # for instance-wise features
        self.parser.add_argument('--no_instance', action='store_false', help='if specified, do *not* add instance map as input')       # 如果指定为True，则不添加实例映射作为输入
        self.parser.add_argument('--instance_feat', action='store_true', help='if specified, add encoded instance features as input')  # 如果指定，添加编码的实例特性作为输入
        self.parser.add_argument('--label_feat', action='store_true', help='if specified, add encoded label features as input')        # 如果指定，添加编码的标签特性作为输入
        self.parser.add_argument('--feat_num', type=int, default=3, help='vector length for encoded features')                         # 编码特征的向量长度
        self.parser.add_argument('--load_features', action='store_true', help='if specified, load precomputed feature maps')           # 如果指定，则加载预计算的特征映射
        self.parser.add_argument('--n_downsample_E', type=int, default=4, help='# of downsampling layers in encoder') 
        self.parser.add_argument('--nef', type=int, default=16, help='# of encoder filters in the first conv layer')        
        self.parser.add_argument('--n_clusters', type=int, default=10, help='number of clusters for features')        

        self.initialized = True

    def parse(self, save=True):
        if not self.initialized:
            self.initialize()
        self.opt = self.parser.parse_args()
        self.opt.isTrain = self.isTrain   # train or test

        str_ids = self.opt.gpu_ids.split(',')
        self.opt.gpu_ids = []
        for str_id in str_ids:
            id = int(str_id)
            if id >= 0:
                self.opt.gpu_ids.append(id)
        
        # set gpu ids
        if len(self.opt.gpu_ids) > 0:
            torch.cuda.set_device(self.opt.gpu_ids[0])

        args = vars(self.opt)

        print('------------ Options -------------')
        for k, v in sorted(args.items()):
            print('%s: %s' % (str(k), str(v)))
        print('-------------- End ----------------')

        # save to the disk        
        expr_dir = os.path.join(self.opt.checkpoints_dir, self.opt.name)
        util.mkdirs(expr_dir)
        if save and not self.opt.continue_train:
            file_name = os.path.join(expr_dir, 'opt.txt')
            with open(file_name, 'wt') as opt_file:
                opt_file.write('------------ Options -------------\n')
                for k, v in sorted(args.items()):
                    opt_file.write('%s: %s\n' % (str(k), str(v)))
                opt_file.write('-------------- End ----------------\n')
        return self.opt

train_options.py

from .base_options import BaseOptions


# train参数选项
class TrainOptions(BaseOptions):
    def initialize(self):
        BaseOptions.initialize(self)
        # for displays
        self.parser.add_argument('--display_freq', type=int, default=100, help='frequency of showing training results on screen')
        self.parser.add_argument('--print_freq', type=int, default=100, help='frequency of showing training results on console')
        self.parser.add_argument('--save_latest_freq', type=int, default=1000, help='frequency of saving the latest results')
        self.parser.add_argument('--save_epoch_freq', type=int, default=10, help='frequency of saving checkpoints at the end of epochs')  # 保存模型的频率为10，每10个epoch保存model参数一次
        self.parser.add_argument('--no_html', action='store_true', help='do not save intermediate training results to [opt.checkpoints_dir]/[opt.name]/web/')
        self.parser.add_argument('--debug', action='store_true', help='only do one epoch and displays at each iteration')

        # for training
        self.parser.add_argument('--continue_train', action='store_true', help='continue training: load the latest model')
        self.parser.add_argument('--load_pretrain', type=str, default='', help='load the pretrained model from the specified location')
        self.parser.add_argument('--which_epoch', type=str, default='latest', help='which epoch to load? set to latest to use latest cached model')  # 哪一个循环加载？默认设置为latest以使用最新的缓存模型
        self.parser.add_argument('--phase', type=str, default='train', help='train, val, test, etc')                    # 默认路径选择为Train数据集
        self.parser.add_argument('--niter', type=int, default=100, help='# of iter at starting learning rate')          # iter的初始学习速率
        self.parser.add_argument('--niter_decay', type=int, default=100, help='# of iter to linearly decay learning rate to zero')
        self.parser.add_argument('--beta1', type=float, default=0.5, help='momentum term of adam')
        self.parser.add_argument('--lr', type=float, default=0.0002, help='initial learning rate for adam')

        # for discriminators        
        self.parser.add_argument('--num_D', type=int, default=2, help='number of discriminators to use')                # 默认判别器数量为2
        self.parser.add_argument('--n_layers_D', type=int, default=3, help='only used if which_model_netD==n_layers')
        self.parser.add_argument('--ndf', type=int, default=64, help='# of discrim filters in first conv layer')    
        self.parser.add_argument('--lambda_feat', type=float, default=10.0, help='weight for feature matching loss')                
        self.parser.add_argument('--no_ganFeat_loss', action='store_true', help='if specified, do *not* use discriminator feature matching loss')  # 如果指定，不使用鉴别器特征匹配丢失
        self.parser.add_argument('--no_vgg_loss', action='store_true', help='if specified, do *not* use VGG feature matching loss')     # 如果指定，不要使用VGG功能匹配丢失
        self.parser.add_argument('--no_lsgan', action='store_true', help='do *not* use least square GAN, if false, use vanilla GAN')    # 不要使用最小二乘GAN，如果为False则使用vanilla GAN
        self.parser.add_argument('--pool_size', type=int, default=0, help='the size of image buffer that stores previously generated images')

        self.isTrain = True                                                                                             # 此处是训练模型，因此若正向传播则isTrain=True；若仅反向传播则isTrain=False

test_options.py

from .base_options import BaseOptions

class TestOptions(BaseOptions):
    def initialize(self):
        BaseOptions.initialize(self)
        self.parser.add_argument('--ntest', type=int, default=float("inf"), help='# of test examples.')
        self.parser.add_argument('--results_dir', type=str, default='./results/', help='saves results here.')
        self.parser.add_argument('--aspect_ratio', type=float, default=1.0, help='aspect ratio of result images')
        self.parser.add_argument('--phase', type=str, default='test', help='train, val, test, etc')
        self.parser.add_argument('--which_epoch', type=str, default='latest', help='which epoch to load? set to latest to use latest cached model')
        self.parser.add_argument('--how_many', type=int, default=39, help='how many test images to run')
        self.parser.add_argument('--cluster_path', type=str, default='features_clustered_010.npy', help='the path for clustered results of encoded features')
        self.parser.add_argument('--use_encoded_image', action='store_true', help='if specified, encode the real image to get the feature map')
        self.parser.add_argument("--export_onnx", type=str, help="export ONNX model to a given file")
        self.parser.add_argument("--engine", type=str, help="run serialized TRT engine")
        self.parser.add_argument("--onnx", type=str, help="run ONNX model via TRT")        
        self.isTrain = False                                                                                            # 不train,只前向传播