def __init__(self, load_weights=False):
super(ASPP, self).__init__()
self.seen = 0
self.frontend_feat = [64, 64, 'M', 128, 128, 'M', 256, 256, 256, 'M', 512, 512, 512]
self.backend_feat = [512, 512, 512, 256, 128, 64]
self.frontend = make_layers(self.frontend_feat)
self.backend = make_layers(self.backend_feat, in_channels=512, dilation=True)
self.output_layer = nn.Conv2d(64, 1, kernel_size=1)
if not load_weights:
mod = models.vgg16(pretrained = True)
utils.weights_normal_init(self)
for i in range(len(self.frontend.state_dict().items())):
list(self.frontend.state_dict().items())[i][1].data[:] = list(mod.state_dict().items())[i][1].data[:]
def __init__(self):
super(PaDNet, self).__init__()
self.frontend_feat = [64, 64, 'M', 128, 128, 'M', 256, 256, 256, 'M', 512, 512, 512]
self.backend_feat = [512, 512, 512, 256, 128, 64]
self.frontend = make_layers(self.frontend_feat)
self.backend_1 = make_layers(self.backend_feat, in_channels=512, d_rate=1)
self.backend_2 = make_layers(self.backend_feat, in_channels=512, d_rate=2)
self.backend_3 = make_layers(self.backend_feat, in_channels=512, d_rate=3)
self.output_layer = nn.Conv2d(64, 1, kernel_size=1)
self.out_layer = nn.Conv2d(3 ,1, kernel_size=1)
self.srn = SRN()
mod = models.vgg16(pretrained=True)
utils.weights_normal_init(self)
for i in range(len(self.frontend.state_dict().items())):
list(self.frontend.state_dict().items())[i][1].data[:] = list(mod.state_dict().items())[i][1].data[:]
import torch.utils.data
import torch.optim as optim
import warnings
import sys
import math
import numpy as np
from tensorboardX import SummaryWriter
import os
from torchvision import transforms
import time
import datetime
warnings.filterwarnings("ignore")
DEVICE = "cuda:0" if torch.cuda.is_available() else "cpu"
models = {
'mcnn': utils.weights_normal_init(mcnn.MCNN(bn=False), dev=0.01),
'csr_net': csr_net.CSRNet(),
'sa_net': sa_net.SANet(input_channels=3,
kernel_size=[1, 3, 5, 7],
bias=True),
'tdf_net': utils.weights_normal_init(tdf_net.TDFNet(), dev=0.01),
'pad_net': pad_net.PaDNet(),
'vgg': vgg.VGG(),
'cbam_net': cbam_net.CBAMNet(),
'big_net': big_net.BIGNet(),
'adcrowd_net': adcrowd_net.ADCrowdNet()
}
def _load_dataset(dataset, zoom_size=4, transform=None):
train_loader = None