def __init__(self, dir, txt_file):
super(Celeb, self).__init__(dir)
self.transform = transforms.Compose([
transforms.RandomCrop((178, 178)),
transforms.Resize(128),
transforms.toTensor()
])
self.dir = dir
self.labels = {}
with open(txt_file) as f:
lines = f.readlines()
for line in lines:
name, label = line.split("\t")
self.labels[name] = eval(label)
def __init__(self, data_dir):
self.data_dir = data_dir
self.label_dict = read_train(self.data_dir)
self.data = []
self.label = []
for key in self.label_dict.keys():
img_path = list(
paths.list_images(os.join(self.data_dir, key) + '/'))
for ip in img_path:
self.data.append(ip)
self.label.append(label_dict[key])
self.transform = transforms.Compose([
transforms.Resize(224),
transforms.RandomHorizontalFlip(),
transforms.RandomVerticalFlip(),
transforms.RandomRotation(45),
transforms.toTensor()
])
def get_views(self, synset_id, model_id):
"""
Retrieves multiple RGB views for a single object.
Arguments:
synset_id (string): eight-digit zero-padded code corresponding to class label for an object
model_id (string): longer name consisting of alphanumeric characters [0-9, a-f]
Example:
synset_id = '02691156'
model_id = '10155655850468db78d106ce0a280f87'
return:
list of all view images for each object in torch.Tensor format
"""
all_views = []
toTensor = transforms.toTensor()
for path in Path(os.path.join(self.root_dir, synset_id, model_id)).glob('**/*.jpg'):
img = toTensor(Image.open(path))
all_views.append(img)
return all_views
import torch
import torchvision
import torchvision.transforms as T
#------------------------------------------------------------------------------------------------------------------------------------------
#STEP1: Loading dataset with suitable transforms
#general transformation to be applied to all images
transform=T.compose([T.toTensor(),T.Normalize((0.5,0.5,0.5),(0.5,0.5,0.5))])
#Get the training dataset apply the transforms and apply load it in a variable with appt. batch size shuffling etc.
trainset=torchvision.datasets.CIFAR10(root='./data',train=True,download=True,transform=transform)
trainloader=torch.utils.data.DataLoader(trainset,batch_size = 4,shuffle=True, num_workers=2)
#Get the training dataset apply the transforms and apply load it in a variable with appt. batch size shuffling etc.
testset=torchvision.datasets.CIFAR10(root='./data',train=False,download=True,transofrm=transform)
testloader=torch.utils.data.Dataloader(testset,shuffle=True,batch_size=4,num_workers=2)
classes=('plane' , 'car', 'bird' , 'cat', 'deer','dog', 'frog', 'horse', 'ship' , 'truck')
#------------------------------------------------------------------------------------------------------------------------------------------
#STEP2:Displaying Training Images
def __init__(self, size, interpolation=Image.BILINEAR):
self.size = size
self.interpolation = interpolation
self.toTensor = transforms.toTensor()
category_size = 10 # one-hot vector
temperature_min = 0.5
ANNEAL_RATE = 0.00003
opt = parser.parse_args()
device = torch.device(opt.device)
model = GumbelVAE(784, latent_size, category_size, device)
optimizer = optim.Adam(model.parameters(), lr=0.001)
kwargs = {'num_workers': 2, 'pin_memory': True}
train_loader = torch.utils.data.DataLoader(datasets.MNIST(
'./data', trian=True, download=True, transform=transforms.toTensor()),
batch_size=opt.batch_size,
shuffle=True,
**kwargs)
test_loader = torch.utils.data.DataLoader(datasets.MNIST(
'./data', train=False, transform=transforms.toTensor()),
batch_size=opt.batch_size,
shuffle=True,
**kwargs)
# reconstruction + KL divergence losses summed over all elements and batch
def loss_function(recon_output, input, encode_output):
BCE = F.binary_cross_entropy(recon_output,
input.view(-1, 784),