File size: 8,486 Bytes
caa56d6 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 |
'''
# author: Zhiyuan Yan
# email: [email protected]
# date: 2023-0706
The code is mainly modified from GitHub link below:
https://github.com/ondyari/FaceForensics/blob/master/classification/network/xception.py
'''
import os
import argparse
import logging
import math
import torch
# import pretrainedmodels
import torch.nn as nn
import torch.nn.functional as F
import torch.utils.model_zoo as model_zoo
from torch.nn import init
from typing import Union
from metrics.registry import BACKBONE
logger = logging.getLogger(__name__)
class SeparableConv2d(nn.Module):
def __init__(self, in_channels, out_channels, kernel_size=1, stride=1, padding=0, dilation=1, bias=False):
super(SeparableConv2d, self).__init__()
self.conv1 = nn.Conv2d(in_channels, in_channels, kernel_size,
stride, padding, dilation, groups=in_channels, bias=bias)
self.pointwise = nn.Conv2d(
in_channels, out_channels, 1, 1, 0, 1, 1, bias=bias)
def forward(self, x):
x = self.conv1(x)
x = self.pointwise(x)
return x
class Block(nn.Module):
def __init__(self, in_filters, out_filters, reps, strides=1, start_with_relu=True, grow_first=True):
super(Block, self).__init__()
if out_filters != in_filters or strides != 1:
self.skip = nn.Conv2d(in_filters, out_filters,
1, stride=strides, bias=False)
self.skipbn = nn.BatchNorm2d(out_filters)
else:
self.skip = None
self.relu = nn.ReLU(inplace=True)
rep = []
filters = in_filters
if grow_first: # whether the number of filters grows first
rep.append(self.relu)
rep.append(SeparableConv2d(in_filters, out_filters,
3, stride=1, padding=1, bias=False))
rep.append(nn.BatchNorm2d(out_filters))
filters = out_filters
for i in range(reps-1):
rep.append(self.relu)
rep.append(SeparableConv2d(filters, filters,
3, stride=1, padding=1, bias=False))
rep.append(nn.BatchNorm2d(filters))
if not grow_first:
rep.append(self.relu)
rep.append(SeparableConv2d(in_filters, out_filters,
3, stride=1, padding=1, bias=False))
rep.append(nn.BatchNorm2d(out_filters))
if not start_with_relu:
rep = rep[1:]
else:
rep[0] = nn.ReLU(inplace=False)
if strides != 1:
rep.append(nn.MaxPool2d(3, strides, 1))
self.rep = nn.Sequential(*rep)
def forward(self, inp):
x = self.rep(inp)
if self.skip is not None:
skip = self.skip(inp)
skip = self.skipbn(skip)
else:
skip = inp
x += skip
return x
def add_gaussian_noise(ins, mean=0, stddev=0.2):
noise = ins.data.new(ins.size()).normal_(mean, stddev)
return ins + noise
@BACKBONE.register_module(module_name="xception")
class Xception(nn.Module):
"""
Xception optimized for the ImageNet dataset, as specified in
https://arxiv.org/pdf/1610.02357.pdf
"""
def __init__(self, xception_config):
""" Constructor
Args:
xception_config: configuration file with the dict format
"""
super(Xception, self).__init__()
self.num_classes = xception_config["num_classes"]
self.mode = xception_config["mode"]
inc = xception_config["inc"]
dropout = xception_config["dropout"]
# Entry flow
self.conv1 = nn.Conv2d(inc, 32, 3, 2, 0, bias=False)
self.bn1 = nn.BatchNorm2d(32)
self.relu = nn.ReLU(inplace=True)
self.conv2 = nn.Conv2d(32, 64, 3, bias=False)
self.bn2 = nn.BatchNorm2d(64)
# do relu here
self.block1 = Block(
64, 128, 2, 2, start_with_relu=False, grow_first=True)
self.block2 = Block(
128, 256, 2, 2, start_with_relu=True, grow_first=True)
self.block3 = Block(
256, 728, 2, 2, start_with_relu=True, grow_first=True)
# middle flow
self.block4 = Block(
728, 728, 3, 1, start_with_relu=True, grow_first=True)
self.block5 = Block(
728, 728, 3, 1, start_with_relu=True, grow_first=True)
self.block6 = Block(
728, 728, 3, 1, start_with_relu=True, grow_first=True)
self.block7 = Block(
728, 728, 3, 1, start_with_relu=True, grow_first=True)
self.block8 = Block(
728, 728, 3, 1, start_with_relu=True, grow_first=True)
self.block9 = Block(
728, 728, 3, 1, start_with_relu=True, grow_first=True)
self.block10 = Block(
728, 728, 3, 1, start_with_relu=True, grow_first=True)
self.block11 = Block(
728, 728, 3, 1, start_with_relu=True, grow_first=True)
# Exit flow
self.block12 = Block(
728, 1024, 2, 2, start_with_relu=True, grow_first=False)
self.conv3 = SeparableConv2d(1024, 1536, 3, 1, 1)
self.bn3 = nn.BatchNorm2d(1536)
# do relu here
self.conv4 = SeparableConv2d(1536, 2048, 3, 1, 1)
self.bn4 = nn.BatchNorm2d(2048)
# used for iid
final_channel = 2048
if self.mode == 'adjust_channel_iid':
final_channel = 512
self.mode = 'adjust_channel'
self.last_linear = nn.Linear(final_channel, self.num_classes)
if dropout:
self.last_linear = nn.Sequential(
nn.Dropout(p=dropout),
nn.Linear(final_channel, self.num_classes)
)
self.adjust_channel = nn.Sequential(
nn.Conv2d(2048, 512, 1, 1),
nn.BatchNorm2d(512),
nn.ReLU(inplace=False),
)
def fea_part1_0(self, x):
x = self.conv1(x)
x = self.bn1(x)
x = self.relu(x)
return x
def fea_part1_1(self, x):
x = self.conv2(x)
x = self.bn2(x)
x = self.relu(x)
return x
def fea_part1(self, x):
x = self.conv1(x)
x = self.bn1(x)
x = self.relu(x)
x = self.conv2(x)
x = self.bn2(x)
x = self.relu(x)
return x
def fea_part2(self, x):
x = self.block1(x)
x = self.block2(x)
x = self.block3(x)
return x
def fea_part3(self, x):
if self.mode == "shallow_xception":
return x
else:
x = self.block4(x)
x = self.block5(x)
x = self.block6(x)
x = self.block7(x)
return x
def fea_part4(self, x):
if self.mode == "shallow_xception":
x = self.block12(x)
else:
x = self.block8(x)
x = self.block9(x)
x = self.block10(x)
x = self.block11(x)
x = self.block12(x)
return x
def fea_part5(self, x):
x = self.conv3(x)
x = self.bn3(x)
x = self.relu(x)
x = self.conv4(x)
x = self.bn4(x)
return x
def features(self, input):
x = self.fea_part1(input)
x = self.fea_part2(x)
x = self.fea_part3(x)
x = self.fea_part4(x)
x = self.fea_part5(x)
if self.mode == 'adjust_channel':
x = self.adjust_channel(x)
return x
def classifier(self, features,id_feat=None):
# for iid
if self.mode == 'adjust_channel':
x = features
else:
x = self.relu(features)
if len(x.shape) == 4:
x = F.adaptive_avg_pool2d(x, (1, 1))
x = x.view(x.size(0), -1)
self.last_emb = x
# for iid
if id_feat!=None:
out = self.last_linear(x-id_feat)
else:
out = self.last_linear(x)
return out
def forward(self, input):
x = self.features(input)
out = self.classifier(x)
return out, x
|