【DeepDream】，代码简单记录

 参考源代码GitHub：

Deep-Dream/resnet.py at master · L1aoXingyu/Deep-Dream · GitHub

但是如果运行里面的代码的话，会报错：

TypeError: __init__() takes from 3 to 5 positional arguments but 9 were given

所以改了一下，让它能运行了。（其实就是重写了ResNet50网络结构，让layers.append(block(self.inplanes, planes, stride, downsample))，只包含四项，而非九项） 

uilt

from PIL import Image


# 使图片大小保持一致,等比缩放
def keep_image_size_open(path, size=(256, 256)):
    img = Image.open(path)  # 读取图片
    temp = max(img.size)  # 取最长边
    mask = Image.new('RGB', (temp, temp), (0, 0, 0))  # mask掩码,全黑
    mask.paste(img, (0, 0))  # 从原点开始粘贴
    mask = mask.resize(size)
    return mask

ResNet

import torch
from torch import nn
from torchvision import models
import torch.utils.model_zoo as model_zoo


model_urls = {
    'resnet18': 'https://download.pytorch.org/models/resnet18-5c106cde.pth',
    'resnet34': 'https://download.pytorch.org/models/resnet34-333f7ec4.pth',
    'resnet50': 'https://download.pytorch.org/models/resnet50-19c8e357.pth',
    'resnet101': 'https://download.pytorch.org/models/resnet101-5d3b4d8f.pth',
    'resnet152': 'https://download.pytorch.org/models/resnet152-b121ed2d.pth',
}


class Bottleneck(nn.Module):
    expansion = 4

    def __init__(self, inplanes, planes, stride=1, downsample=None):
        super(Bottleneck, self).__init__()
        self.conv1 = nn.Conv2d(inplanes, planes, kernel_size=1, bias=False)
        self.bn1 = nn.BatchNorm2d(planes)
        self.conv2 = nn.Conv2d(planes, planes, kernel_size=3, stride=stride,
                               padding=1, bias=False)
        self.bn2 = nn.BatchNorm2d(planes)
        self.conv3 = nn.Conv2d(planes, planes * 4, kernel_size=1, bias=False)
        self.bn3 = nn.BatchNorm2d(planes * 4)
        self.relu = nn.ReLU(inplace=True)
        self.downsample = downsample
        self.stride = stride

    def forward(self, x):
        residual = x

        out = self.conv1(x)
        out = self.bn1(out)
        out = self.relu(out)

        out = self.conv2(out)
        out = self.bn2(out)
        out = self.relu(out)

        out = self.conv3(out)
        out = self.bn3(out)

        if self.downsample is not None:
            residual = self.downsample(x)

        out += residual
        out = self.relu(out)

        return out


class ResNet(nn.Module):

    def __init__(self, block, layers, num_classes=1000):
        self.inplanes = 64
        super(ResNet, self).__init__()
        self.conv1 = nn.Conv2d(3, 64, kernel_size=7, stride=2, padding=3,
                               bias=False)
        self.bn1 = nn.BatchNorm2d(64)
        self.relu = nn.ReLU(inplace=True)
        self.maxpool = nn.MaxPool2d(kernel_size=3, stride=2, padding=1)
        self.layer1 = self._make_layer(block, 64, layers[0])
        self.layer2 = self._make_layer(block, 128, layers[1], stride=2)
        self.layer3 = self._make_layer(block, 256, layers[2], stride=2)
        self.layer4 = self._make_layer(block, 512, layers[3], stride=2)
        self.avgpool = nn.AvgPool2d(7, stride=1)
        self.fc = nn.Linear(512 * block.expansion, num_classes)

        for m in self.modules():
            if isinstance(m, nn.Conv2d):
                nn.init.kaiming_normal_(m.weight, mode='fan_out', nonlinearity='relu')
            elif isinstance(m, nn.BatchNorm2d):
                nn.init.constant_(m.weight, 1)
                nn.init.constant_(m.bias, 0)

    def _make_layer(self, block, planes, blocks, stride=1):
        downsample = None
        if stride != 1 or self.inplanes != planes * block.expansion:
            downsample = nn.Sequential(
                nn.Conv2d(self.inplanes, planes * block.expansion,
                          kernel_size=1, stride=stride, bias=False),
                nn.BatchNorm2d(planes * block.expansion),
            )

        layers = []
        layers.append(block(self.inplanes, planes, stride, downsample))
        self.inplanes = planes * block.expansion
        for i in range(1, blocks):
            layers.append(block(self.inplanes, planes))

        return nn.Sequential(*layers)

    def forward(self, x):
        x = self.conv1(x)
        x = self.bn1(x)
        x = self.relu(x)
        x = self.maxpool(x)

        x = self.layer1(x)
        x = self.layer2(x)
        x = self.layer3(x)
        x = self.layer4(x)

        x = self.avgpool(x)
        x = x.view(x.size(0), -1)
        x = self.fc(x)

        return x

class CustomResNet(ResNet):  ##models.resnet.ResNet):
    def forward(self, x, end_layer):
        """
        end_layer range from 1 to 4
        """
        x = self.conv1(x)
        x = self.bn1(x)
        x = self.relu(x)
        x = self.maxpool(x)

        layers = [self.layer1, self.layer2, self.layer3, self.layer4]
        for i in range(end_layer):
            x = layers[i](x)
        return x


def resnet50(pretrained=False, **kwargs):
    model = CustomResNet(Bottleneck, [3, 4, 6, 3], **kwargs)
    if pretrained:
        model.load_state_dict(model_zoo.load_url(model_urls['resnet50']))
    return model

if __name__ == '__main__':
    model = resnet50(pretrained=False)

deepdream

import numpy as np
import torch
from util import showtensor
import scipy.ndimage as nd
from torch.autograd import Variable


def objective_L2(dst, guide_features):
    return dst.data


def make_step(img, model, control=None, distance=objective_L2):
    mean = np.array([0.485, 0.456, 0.406]).reshape([3, 1, 1])
    std = np.array([0.229, 0.224, 0.225]).reshape([3, 1, 1])

    learning_rate = 2e-2
    max_jitter = 32
    num_iterations = 20
    show_every = 10
    end_layer = 3
    guide_features = control

    for i in range(num_iterations):
        shift_x, shift_y = np.random.randint(-max_jitter, max_jitter + 1, 2)
        img = np.roll(np.roll(img, shift_x, -1), shift_y, -2)
        # apply jitter shift
        model.zero_grad()
        img_tensor = torch.Tensor(img)
        if torch.cuda.is_available():
            img_variable = Variable(img_tensor.cuda(), requires_grad=True)
        else:
            img_variable = Variable(img_tensor, requires_grad=True)

        act_value = model.forward(img_variable, end_layer)
        diff_out = distance(act_value, guide_features)
        act_value.backward(diff_out)
        ratio = np.abs(img_variable.grad.data.cpu().numpy()).mean()
        learning_rate_use = learning_rate / ratio
        img_variable.data.add_(img_variable.grad.data * learning_rate_use)

        img_variable = img_variable.clamp(0, 255)

        img = img_variable.data.cpu().numpy()  # b, c, h, w
        img = np.roll(np.roll(img, -shift_x, -1), -shift_y, -2)
        img[0, :, :, :] = np.clip(img[0, :, :, :], -mean / std,
                                  (1 - mean) / std)
        if i == 0 or (i + 1) % show_every == 0:
            showtensor(img)
    return img


def dream(model,
          base_img,
          octave_n=6,
          octave_scale=1.4,
          control=None,
          distance=objective_L2):
    octaves = [base_img]
    for i in range(octave_n - 1):
        octaves.append(
            nd.zoom(
                octaves[-1], (1, 1, 1.0 / octave_scale, 1.0 / octave_scale),
                order=1))

    detail = np.zeros_like(octaves[-1])
    for octave, octave_base in enumerate(octaves[::-1]):
        h, w = octave_base.shape[-2:]
        if octave > 0:
            h1, w1 = detail.shape[-2:]
            detail = nd.zoom(
                detail, (1, 1, 1.0 * h / h1, 1.0 * w / w1), order=1)

        input_oct = octave_base + detail
        print(input_oct.shape)
        out = make_step(input_oct, model, control, distance=distance)
        detail = out - octave_base
        return out

train

## 定义一些参数

device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')

img_transform = transforms.Compose([
    transforms.ToTensor(),
    transforms.Normalize([0.5,0.5,0.5],[0.5,0.5,0.5])
])

input_img = Image.open('./sky.jpg')
# input_img = cv2.imread('./cat.png')
print(input_img.size)
input_tensor = img_transform(input_img).unsqueeze(0)    ## 输出的是【1,3,224,224】
input_np = input_tensor.numpy()

## 加载模型
model = resnet50(pretrained=True).to(device)   ## True加载参数
for param in model.parameters():
    param.requires_grad = False
a = dream(model, input_np)
# a = a.clamp(0, 1)
# print('!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!')
# print(a.shape)
decode_img = np.array((a))
decode_img = decode_img.squeeze(0)
decode_img = decode_img.transpose((1, 2, 0))

print(decode_img)
plt.imshow(decode_img) # 生成图片 3
plt.show()

效果图

 源图：