def optimize(args):
""" Gatys et al. CVPR 2017
ref: Image Style Transfer Using Convolutional Neural Networks
"""
# load the content and style target
content_image = utils.tensor_load_rgbimage(args.content_image,
size=args.content_size,
keep_asp=True)
content_image = content_image.unsqueeze(0)
content_image = Variable(utils.preprocess_batch(content_image),
requires_grad=False)
content_image = utils.subtract_imagenet_mean_batch(content_image)
style_image = utils.tensor_load_rgbimage(args.style_image,
size=args.style_size)
style_image = style_image.unsqueeze(0)
style_image = Variable(utils.preprocess_batch(style_image),
requires_grad=False)
style_image = utils.subtract_imagenet_mean_batch(style_image)
# load the pre-trained vgg-16 and extract features
vgg = Vgg16()
utils.init_vgg16(args.vgg_model_dir)
vgg.load_state_dict(
torch.load(os.path.join(args.vgg_model_dir, "vgg16.weight")))
if args.cuda:
content_image = content_image.cuda()
style_image = style_image.cuda()
vgg.cuda()
features_content = vgg(content_image)
f_xc_c = Variable(features_content[1].data, requires_grad=False)
features_style = vgg(style_image)
gram_style = [utils.gram_matrix(y) for y in features_style]
# init optimizer
output = Variable(content_image.data, requires_grad=True)
optimizer = Adam([output], lr=args.lr)
mse_loss = torch.nn.MSELoss()
# optimizing the images
for e in range(args.iters):
utils.imagenet_clamp_batch(output, 0, 255)
optimizer.zero_grad()
features_y = vgg(output)
content_loss = args.content_weight * mse_loss(features_y[1], f_xc_c)
style_loss = 0.
for m in range(len(features_y)):
gram_y = utils.gram_matrix(features_y[m])
gram_s = Variable(gram_style[m].data, requires_grad=False)
style_loss += args.style_weight * mse_loss(gram_y, gram_s)
total_loss = content_loss + style_loss
if (e + 1) % args.log_interval == 0:
print(total_loss.data.cpu().numpy()[0])
total_loss.backward()
optimizer.step()
# save the image
output = utils.add_imagenet_mean_batch(output)
utils.tensor_save_bgrimage(output.data[0], args.output_image, args.cuda)
def evaluate(args):
content_image = utils.tensor_load_rgbimage(args.content_image,
size=args.content_size,
keep_asp=True)
content_image = content_image.unsqueeze(0)
style = utils.tensor_load_rgbimage(args.style_image, size=args.style_size)
style = style.unsqueeze(0)
style = utils.preprocess_batch(style)
vgg = Vgg16()
utils.init_vgg16(args.vgg_model_dir)
vgg.load_state_dict(
torch.load(os.path.join(args.vgg_model_dir, "vgg16.weight")))
style_model = Net()
style_model.load_state_dict(torch.load(args.model))
if args.cuda:
style_model.cuda()
vgg.cuda()
content_image = content_image.cuda()
style = style.cuda()
style_v = Variable(style, volatile=True)
style_v = utils.subtract_imagenet_mean_batch(style_v)
features_style = vgg(style_v)
gram_style = [utils.gram_matrix(y) for y in features_style]
content_image = Variable(utils.preprocess_batch(content_image),
volatile=True)
style_model.setTarget(gram_style[2].data)
output = style_model(content_image)
utils.tensor_save_bgrimage(output.data[0], args.output_image, args.cuda)
def extract_feats(args):
vgg = Vgg16()
utils.init_vgg16(args.vgg_model_dir)
cap = dset.CocoCaptions(
root='/Pulsar1/Datasets/coco/train2014/train2014',
annFile=
'/Neutron9/sahil.c/datasets/annotations/captions_train2014.json',
transform=transforms.ToTensor())
print('Number of samples: ', len(cap))
for i, t in cap:
image = i.unsqueeze(0)
image = Variable(utils.preprocess_batch(image), requires_grad=False)
image = utils.subtract_imagenet_mean_batch(image)
features_content = vgg(image)
def train(args):
check_paths(args)
np.random.seed(args.seed)
torch.manual_seed(args.seed)
if args.cuda:
torch.cuda.manual_seed(args.seed)
kwargs = {'num_workers': 0, 'pin_memory': False}
else:
kwargs = {}
transform = transforms.Compose([
transforms.Scale(args.image_size),
transforms.CenterCrop(args.image_size),
transforms.ToTensor(),
transforms.Lambda(lambda x: x.mul(255))
])
train_dataset = datasets.ImageFolder(args.dataset, transform)
train_loader = DataLoader(train_dataset,
batch_size=args.batch_size,
**kwargs)
style_model = Net(ngf=args.ngf)
if args.resume is not None:
print('Resuming, initializing using weight from {}.'.format(
args.resume))
style_model.load_state_dict(torch.load(args.resume))
print(style_model)
optimizer = Adam(style_model.parameters(), args.lr)
mse_loss = torch.nn.MSELoss()
vgg = Vgg16()
utils.init_vgg16(args.vgg_model_dir)
vgg.load_state_dict(
torch.load(os.path.join(args.vgg_model_dir, "vgg16.weight")))
if args.cuda:
style_model.cuda()
vgg.cuda()
style_loader = StyleLoader(args.style_folder, args.style_size)
for e in range(args.epochs):
style_model.train()
agg_content_loss = 0.
agg_style_loss = 0.
count = 0
for batch_id, (x, _) in enumerate(train_loader):
n_batch = len(x)
count += n_batch
optimizer.zero_grad()
x = Variable(utils.preprocess_batch(x))
if args.cuda:
x = x.cuda()
style_v = style_loader.get(batch_id)
style_model.setTarget(style_v)
style_v = utils.subtract_imagenet_mean_batch(style_v)
features_style = vgg(style_v)
gram_style = [utils.gram_matrix(y) for y in features_style]
y = style_model(x)
xc = Variable(x.data.clone(), volatile=True)
y = utils.subtract_imagenet_mean_batch(y)
xc = utils.subtract_imagenet_mean_batch(xc)
features_y = vgg(y)
features_xc = vgg(xc)
f_xc_c = Variable(features_xc[1].data, requires_grad=False)
content_loss = args.content_weight * mse_loss(
features_y[1], f_xc_c)
style_loss = 0.
for m in range(len(features_y)):
gram_y = utils.gram_matrix(features_y[m])
gram_s = Variable(gram_style[m].data,
requires_grad=False).repeat(
args.batch_size, 1, 1, 1)
style_loss += args.style_weight * mse_loss(
gram_y, gram_s[:n_batch, :, :])
total_loss = content_loss + style_loss
total_loss.backward()
optimizer.step()
agg_content_loss += content_loss.data[0]
agg_style_loss += style_loss.data[0]
if (batch_id + 1) % args.log_interval == 0:
mesg = "{}\tEpoch {}:\t[{}/{}]\tcontent: {:.6f}\tstyle: {:.6f}\ttotal: {:.6f}".format(
time.ctime(), e + 1, count, len(train_dataset),
agg_content_loss / (batch_id + 1),
agg_style_loss / (batch_id + 1),
(agg_content_loss + agg_style_loss) / (batch_id + 1))
print(mesg)
if (batch_id + 1) % (4 * args.log_interval) == 0:
# save model
style_model.eval()
style_model.cpu()
save_model_filename = "Epoch_" + str(e) + "iters_" + str(
count) + "_" + str(time.ctime()).replace(
' ', '_') + "_" + str(args.content_weight) + "_" + str(
args.style_weight) + ".model"
save_model_path = os.path.join(args.save_model_dir,
save_model_filename)
torch.save(style_model.state_dict(), save_model_path)
style_model.train()
style_model.cuda()
print("\nCheckpoint, trained model saved at", save_model_path)
# save model
style_model.eval()
style_model.cpu()
save_model_filename = "Final_epoch_" + str(args.epochs) + "_" + str(
time.ctime()).replace(' ', '_') + "_" + str(
args.content_weight) + "_" + str(args.style_weight) + ".model"
save_model_path = os.path.join(args.save_model_dir, save_model_filename)
torch.save(style_model.state_dict(), save_model_path)
print("\nDone, trained model saved at", save_model_path)