for id, (data, depth, img_name, img_size) in enumerate(test_loader):
print('testing bach %d' % id)
inputs = Variable(data).cuda()
inputs_depth = Variable(depth).cuda()
n, c, h, w = inputs.size()
depth = inputs_depth.view(n, h, w, 1).repeat(1, 1, 1, c)
depth = depth.transpose(3, 1)
depth = depth.transpose(3, 2)
h1, h2, h3, h4, h5 = model_rgb(inputs) # RGBNet's output
d1, d2, d3, d4, d5 = model_depth(depth) # DepthNet's output
outputs_all = model_fusion(h1, h2, h3, h4, h5, d1, d2, d3, d4,
d5) # Final output
outputs_all = F.softmax(outputs_all, dim=1)
outputs = outputs_all[0][1]
outputs = outputs.cpu().data.resize_(h, w)
imsave(os.path.join(MapRoot, img_name[0] + '.png'), outputs, img_size)
# -------------------------- validation --------------------------- #
torch.cuda.empty_cache()
print("\nevaluating mae....")
F_measure, mae = get_FM(salpath=MapRoot + '/',
gtpath=test_dataRoot + '/test_masks/')
print('F_measure:', F_measure)
print('MAE:', mae)
import torch
from torch.autograd import Variable
from torch.utils.data import DataLoader
import torchvision
import torch.nn.functional as F
import torch.optim as optim
from dataset_loader import MyData, MyTestData, DTestData
from model import FocalNet, FocalNet_sub
from conv_lstm import ConvLSTM
from functions import imsave
import argparse
from Trainer_Teacher import Trainer
import os
if __name__ == '__main__':
configurations = {
1: dict(
max_iteration=500000,
lr=1.0e-10,
momentum=0.99,
weight_decay=0.0005,
spshot=10000,
nclass=2,
sshow=10,
focal_num=12,
)
}
parser=argparse.ArgumentParser()
parser.add_argument('--phase', type=str, default='test', help='train or test')
parser.add_argument('--param', type=str, default=True, help='path to pre-trained parameters')
""" Title: Depth-induced Multi-scale Recurrent Attention Network for Saliency Detection Author: Wei Ji, Jingjing Li E-mail: [email protected] """ import torch from torch.autograd import Variable from torch.utils.data import DataLoader import torchvision import torch.nn.functional as F import torch.optim as optim from dataset_loader import MyData, MyTestData from model import RGBNet,DepthNet from fusion import ConvLSTM from functions import imsave import argparse from trainer import Trainer import os configurations = { # same configuration as original work # https://github.com/shelhamer/fcn.berkeleyvision.org 1: dict( max_iteration=1000000, lr=1.0e-10, momentum=0.99, weight_decay=0.0005, spshot=20000, nclass=2, sshow=10,
inputs = Variable(data).cuda()
n, c, h, w = inputs.size()
begin_time = time.time()
low_1, low_2, high_1, high_2, high_3 = model_rgb(inputs)
Features, _, _, Edge, _, _, Depth, Sal = model_intergration(
low_1, low_2, high_1, high_2, high_3)
outputs = model_att(Features, Edge, Sal, Depth)
outputs = F.softmax(outputs, dim=1)
outputs = outputs[0][1]
outputs = outputs.cpu().data.resize_(h, w)
end_time = time.time()
run_time = end_time - begin_time
res.append(run_time)
imsave(os.path.join(MapRoot, img_name[0] + '.png'), outputs, img_size)
# ---------------- Visual Results ------------------ #
# Edge
out1 = F.softmax(Edge, dim=1)
out1 = out1[0][1]
out1 = out1.cpu().data.resize_(h, w)
imsave(os.path.join(args.out1, img_name[0] + '.png'), out1, img_size)
# Depth
out2 = Depth[0][0]
out2 = out2.cpu().data.resize_(h, w)
imsave(os.path.join(args.out2, img_name[0] + '.png'), out2, img_size)
# Sal-Att
out3 = Sal[0][1]
out3 = out3.cpu().data.resize_(h, w)
imsave(os.path.join(args.out3, img_name[0] + '.png'), out3, img_size)
import torch
from torch.autograd import Variable
from torch.utils.data import DataLoader
import torchvision
import torch.nn.functional as F
import torch.optim as optim
from dataset_loader import MyData, MyTestData
from model import FocalNet, FocalNet_sub
from conv_lstm import ConvLSTM
from functions import imsave
import argparse
from Trainer_Student import Trainer
from resnet_18 import Resnet_18
import os
import imageio
if __name__ == '__main__':
configurations = {
1: dict(
max_iteration=300000,
lr=1.0e-10,
momentum=0.99,
weight_decay=0.0005,
spshot=10000,
nclass=2,
sshow=10,
focal_num=12,
)
}
parser=argparse.ArgumentParser()