def main(args: argparse.Namespace):
logger = CompleteLogger(args.log, args.phase)
print(args)
if args.seed is not None:
random.seed(args.seed)
np.random.seed(args.seed)
torch.manual_seed(args.seed)
cudnn.deterministic = True
warnings.warn('You have chosen to seed training. '
'This will turn on the CUDNN deterministic setting, '
'which can slow down your training considerably! '
'You may see unexpected behavior when restarting '
'from checkpoints.')
cudnn.benchmark = True
# Data loading code
train_transform = utils.get_train_transform(args.height,
args.width,
args.train_resizing,
random_horizontal_flip=True,
random_color_jitter=False,
random_gray_scale=False)
val_transform = utils.get_val_transform(args.height, args.width)
print("train_transform: ", train_transform)
print("val_transform: ", val_transform)
working_dir = osp.dirname(osp.abspath(__file__))
root = osp.join(working_dir, args.root)
# source dataset
source_dataset = datasets.__dict__[args.source](
root=osp.join(root, args.source.lower()))
sampler = RandomDomainMultiInstanceSampler(
source_dataset.train,
batch_size=args.batch_size,
n_domains_per_batch=2,
num_instances=args.num_instances)
train_loader = DataLoader(convert_to_pytorch_dataset(
source_dataset.train,
root=source_dataset.images_dir,
transform=train_transform),
batch_size=args.batch_size,
num_workers=args.workers,
sampler=sampler,
pin_memory=True,
drop_last=True)
train_iter = ForeverDataIterator(train_loader)
val_loader = DataLoader(convert_to_pytorch_dataset(
list(set(source_dataset.query) | set(source_dataset.gallery)),
root=source_dataset.images_dir,
transform=val_transform),
batch_size=args.batch_size,
num_workers=args.workers,
shuffle=False,
pin_memory=True)
# target dataset
target_dataset = datasets.__dict__[args.target](
root=osp.join(root, args.target.lower()))
test_loader = DataLoader(convert_to_pytorch_dataset(
list(set(target_dataset.query) | set(target_dataset.gallery)),
root=target_dataset.images_dir,
transform=val_transform),
batch_size=args.batch_size,
num_workers=args.workers,
shuffle=False,
pin_memory=True)
# create model
num_classes = source_dataset.num_train_pids
backbone = models.__dict__[args.arch](mix_layers=args.mix_layers,
mix_p=args.mix_p,
mix_alpha=args.mix_alpha,
resnet_class=ReidResNet,
pretrained=True)
model = ReIdentifier(backbone, num_classes,
finetune=args.finetune).to(device)
model = DataParallel(model)
# define optimizer and learning rate scheduler
optimizer = Adam(model.module.get_parameters(base_lr=args.lr,
rate=args.rate),
args.lr,
weight_decay=args.weight_decay)
lr_scheduler = WarmupMultiStepLR(optimizer,
args.milestones,
gamma=0.1,
warmup_factor=0.1,
warmup_steps=args.warmup_steps)
# resume from the best checkpoint
if args.phase != 'train':
checkpoint = torch.load(logger.get_checkpoint_path('best'),
map_location='cpu')
model.load_state_dict(checkpoint)
# analysis the model
if args.phase == 'analysis':
# plot t-SNE
utils.visualize_tsne(source_loader=val_loader,
target_loader=test_loader,
model=model,
filename=osp.join(logger.visualize_directory,
'analysis', 'TSNE.pdf'),
device=device)
# visualize ranked results
visualize_ranked_results(test_loader,
model,
target_dataset.query,
target_dataset.gallery,
device,
visualize_dir=logger.visualize_directory,
width=args.width,
height=args.height,
rerank=args.rerank)
return
if args.phase == 'test':
print("Test on source domain:")
validate(val_loader,
model,
source_dataset.query,
source_dataset.gallery,
device,
cmc_flag=True,
rerank=args.rerank)
print("Test on target domain:")
validate(test_loader,
model,
target_dataset.query,
target_dataset.gallery,
device,
cmc_flag=True,
rerank=args.rerank)
return
# define loss function
criterion_ce = CrossEntropyLossWithLabelSmooth(num_classes).to(device)
criterion_triplet = SoftTripletLoss(margin=args.margin).to(device)
# start training
best_val_mAP = 0.
best_test_mAP = 0.
for epoch in range(args.epochs):
# print learning rate
print(lr_scheduler.get_lr())
# train for one epoch
train(train_iter, model, criterion_ce, criterion_triplet, optimizer,
epoch, args)
# update learning rate
lr_scheduler.step()
if (epoch + 1) % args.eval_step == 0 or (epoch == args.epochs - 1):
# evaluate on validation set
print("Validation on source domain...")
_, val_mAP = validate(val_loader,
model,
source_dataset.query,
source_dataset.gallery,
device,
cmc_flag=True)
# remember best mAP and save checkpoint
torch.save(model.state_dict(),
logger.get_checkpoint_path('latest'))
if val_mAP > best_val_mAP:
shutil.copy(logger.get_checkpoint_path('latest'),
logger.get_checkpoint_path('best'))
best_val_mAP = max(val_mAP, best_val_mAP)
# evaluate on test set
print("Test on target domain...")
_, test_mAP = validate(test_loader,
model,
target_dataset.query,
target_dataset.gallery,
device,
cmc_flag=True,
rerank=args.rerank)
best_test_mAP = max(test_mAP, best_test_mAP)
# evaluate on test set
model.load_state_dict(torch.load(logger.get_checkpoint_path('best')))
print("Test on target domain:")
_, test_mAP = validate(test_loader,
model,
target_dataset.query,
target_dataset.gallery,
device,
cmc_flag=True,
rerank=args.rerank)
print("test mAP on target = {}".format(test_mAP))
print("oracle mAP on target = {}".format(best_test_mAP))
logger.close()
def main(args: argparse.Namespace):
logger = CompleteLogger(args.log, args.phase)
print(args)
if args.seed is not None:
random.seed(args.seed)
np.random.seed(args.seed)
torch.manual_seed(args.seed)
cudnn.deterministic = True
warnings.warn('You have chosen to seed training. '
'This will turn on the CUDNN deterministic setting, '
'which can slow down your training considerably! '
'You may see unexpected behavior when restarting '
'from checkpoints.')
cudnn.benchmark = True
# Data loading code
train_transform = utils.get_train_transform(args.height,
args.width,
args.train_resizing,
random_horizontal_flip=True,
random_color_jitter=False,
random_gray_scale=False,
random_erasing=True)
val_transform = utils.get_val_transform(args.height, args.width)
print("train_transform: ", train_transform)
print("val_transform: ", val_transform)
working_dir = osp.dirname(osp.abspath(__file__))
source_root = osp.join(working_dir, args.source_root)
target_root = osp.join(working_dir, args.target_root)
# source dataset
source_dataset = datasets.__dict__[args.source](
root=osp.join(source_root, args.source.lower()))
val_loader = DataLoader(convert_to_pytorch_dataset(
list(set(source_dataset.query) | set(source_dataset.gallery)),
root=source_dataset.images_dir,
transform=val_transform),
batch_size=args.batch_size,
num_workers=args.workers,
shuffle=False,
pin_memory=True)
# target dataset
target_dataset = datasets.__dict__[args.target](
root=osp.join(target_root, args.target.lower()))
cluster_loader = DataLoader(convert_to_pytorch_dataset(
target_dataset.train,
root=target_dataset.images_dir,
transform=val_transform),
batch_size=args.batch_size,
num_workers=args.workers,
shuffle=False,
pin_memory=True)
test_loader = DataLoader(convert_to_pytorch_dataset(
list(set(target_dataset.query) | set(target_dataset.gallery)),
root=target_dataset.images_dir,
transform=val_transform),
batch_size=args.batch_size,
num_workers=args.workers,
shuffle=False,
pin_memory=True)
# create model
num_classes = args.num_clusters
backbone = utils.get_model(args.arch)
pool_layer = nn.Identity() if args.no_pool else None
model = ReIdentifier(backbone,
num_classes,
finetune=args.finetune,
pool_layer=pool_layer).to(device)
model = DataParallel(model)
# load pretrained weights
pretrained_model = torch.load(args.pretrained_model_path)
utils.copy_state_dict(model, pretrained_model)
# resume from the best checkpoint
if args.phase != 'train':
checkpoint = torch.load(logger.get_checkpoint_path('best'),
map_location='cpu')
utils.copy_state_dict(model, checkpoint['model'])
# analysis the model
if args.phase == 'analysis':
# plot t-SNE
utils.visualize_tsne(source_loader=val_loader,
target_loader=test_loader,
model=model,
filename=osp.join(logger.visualize_directory,
'analysis', 'TSNE.pdf'),
device=device)
# visualize ranked results
visualize_ranked_results(test_loader,
model,
target_dataset.query,
target_dataset.gallery,
device,
visualize_dir=logger.visualize_directory,
width=args.width,
height=args.height,
rerank=args.rerank)
return
if args.phase == 'test':
print("Test on Source domain:")
validate(val_loader,
model,
source_dataset.query,
source_dataset.gallery,
device,
cmc_flag=True,
rerank=args.rerank)
print("Test on target domain:")
validate(test_loader,
model,
target_dataset.query,
target_dataset.gallery,
device,
cmc_flag=True,
rerank=args.rerank)
return
# define loss function
criterion_ce = CrossEntropyLossWithLabelSmooth(num_classes).to(device)
criterion_triplet = SoftTripletLoss(margin=args.margin).to(device)
# optionally resume from a checkpoint
if args.resume:
checkpoint = torch.load(args.resume, map_location='cpu')
utils.copy_state_dict(model, checkpoint['model'])
args.start_epoch = checkpoint['epoch'] + 1
# start training
best_test_mAP = 0.
for epoch in range(args.start_epoch, args.epochs):
# run clustering algorithm and generate pseudo labels
if args.clustering_algorithm == 'kmeans':
train_target_iter = run_kmeans(cluster_loader, model,
target_dataset, train_transform,
args)
elif args.clustering_algorithm == 'dbscan':
train_target_iter, num_classes = run_dbscan(
cluster_loader, model, target_dataset, train_transform, args)
# define cross entropy loss with current number of classes
criterion_ce = CrossEntropyLossWithLabelSmooth(num_classes).to(device)
# define optimizer
optimizer = Adam(model.module.get_parameters(base_lr=args.lr,
rate=args.rate),
args.lr,
weight_decay=args.weight_decay)
# train for one epoch
train(train_target_iter, model, optimizer, criterion_ce,
criterion_triplet, epoch, args)
if (epoch + 1) % args.eval_step == 0 or (epoch == args.epochs - 1):
# remember best mAP and save checkpoint
torch.save({
'model': model.state_dict(),
'epoch': epoch
}, logger.get_checkpoint_path(epoch))
print("Test on target domain...")
_, test_mAP = validate(test_loader,
model,
target_dataset.query,
target_dataset.gallery,
device,
cmc_flag=True,
rerank=args.rerank)
if test_mAP > best_test_mAP:
shutil.copy(logger.get_checkpoint_path(epoch),
logger.get_checkpoint_path('best'))
best_test_mAP = max(test_mAP, best_test_mAP)
print("best mAP on target = {}".format(best_test_mAP))
logger.close()
def main(args: argparse.Namespace):
logger = CompleteLogger(args.log, args.phase)
print(args)
if args.seed is not None:
random.seed(args.seed)
np.random.seed(args.seed)
torch.manual_seed(args.seed)
cudnn.deterministic = True
warnings.warn('You have chosen to seed training. '
'This will turn on the CUDNN deterministic setting, '
'which can slow down your training considerably! '
'You may see unexpected behavior when restarting '
'from checkpoints.')
cudnn.benchmark = True
# Data loading code
train_transform = utils.get_train_transform(args.height,
args.width,
args.train_resizing,
random_horizontal_flip=True,
random_color_jitter=False,
random_gray_scale=False,
random_erasing=True)
val_transform = utils.get_val_transform(args.height, args.width)
print("train_transform: ", train_transform)
print("val_transform: ", val_transform)
working_dir = osp.dirname(osp.abspath(__file__))
source_root = osp.join(working_dir, args.source_root)
target_root = osp.join(working_dir, args.target_root)
# source dataset
source_dataset = datasets.__dict__[args.source](
root=osp.join(source_root, args.source.lower()))
sampler = RandomMultipleGallerySampler(source_dataset.train,
args.num_instances)
train_source_loader = DataLoader(convert_to_pytorch_dataset(
source_dataset.train,
root=source_dataset.images_dir,
transform=train_transform),
batch_size=args.batch_size,
num_workers=args.workers,
sampler=sampler,
pin_memory=True,
drop_last=True)
train_source_iter = ForeverDataIterator(train_source_loader)
cluster_source_loader = DataLoader(convert_to_pytorch_dataset(
source_dataset.train,
root=source_dataset.images_dir,
transform=val_transform),
batch_size=args.batch_size,
num_workers=args.workers,
shuffle=False,
pin_memory=True)
val_loader = DataLoader(convert_to_pytorch_dataset(
list(set(source_dataset.query) | set(source_dataset.gallery)),
root=source_dataset.images_dir,
transform=val_transform),
batch_size=args.batch_size,
num_workers=args.workers,
shuffle=False,
pin_memory=True)
# target dataset
target_dataset = datasets.__dict__[args.target](
root=osp.join(target_root, args.target.lower()))
cluster_target_loader = DataLoader(convert_to_pytorch_dataset(
target_dataset.train,
root=target_dataset.images_dir,
transform=val_transform),
batch_size=args.batch_size,
num_workers=args.workers,
shuffle=False,
pin_memory=True)
test_loader = DataLoader(convert_to_pytorch_dataset(
list(set(target_dataset.query) | set(target_dataset.gallery)),
root=target_dataset.images_dir,
transform=val_transform),
batch_size=args.batch_size,
num_workers=args.workers,
shuffle=False,
pin_memory=True)
n_s_classes = source_dataset.num_train_pids
args.n_classes = n_s_classes + len(target_dataset.train)
args.n_s_classes = n_s_classes
args.n_t_classes = len(target_dataset.train)
# create model
backbone = models.__dict__[args.arch](pretrained=True)
pool_layer = nn.Identity() if args.no_pool else None
model = ReIdentifier(backbone,
args.n_classes,
finetune=args.finetune,
pool_layer=pool_layer)
features_dim = model.features_dim
idm_bn_names = filter_layers(args.stage)
convert_dsbn_idm(model, idm_bn_names, idm=False)
model = model.to(device)
model = DataParallel(model)
# resume from the best checkpoint
if args.phase != 'train':
checkpoint = torch.load(logger.get_checkpoint_path('best'),
map_location='cpu')
utils.copy_state_dict(model, checkpoint['model'])
# analysis the model
if args.phase == 'analysis':
# plot t-SNE
utils.visualize_tsne(source_loader=val_loader,
target_loader=test_loader,
model=model,
filename=osp.join(logger.visualize_directory,
'analysis', 'TSNE.pdf'),
device=device)
# visualize ranked results
visualize_ranked_results(test_loader,
model,
target_dataset.query,
target_dataset.gallery,
device,
visualize_dir=logger.visualize_directory,
width=args.width,
height=args.height,
rerank=args.rerank)
return
if args.phase == 'test':
print("Test on target domain:")
validate(test_loader,
model,
target_dataset.query,
target_dataset.gallery,
device,
cmc_flag=True,
rerank=args.rerank)
return
# create XBM
dataset_size = len(source_dataset.train) + len(target_dataset.train)
memory_size = int(args.ratio * dataset_size)
xbm = XBM(memory_size, features_dim)
# initialize source-domain class centroids
source_feature_dict = extract_reid_feature(cluster_source_loader,
model,
device,
normalize=True)
source_features_per_id = {}
for f, pid, _ in source_dataset.train:
if pid not in source_features_per_id:
source_features_per_id[pid] = []
source_features_per_id[pid].append(source_feature_dict[f].unsqueeze(0))
source_centers = [
torch.cat(source_features_per_id[pid], 0).mean(0)
for pid in sorted(source_features_per_id.keys())
]
source_centers = torch.stack(source_centers, 0)
source_centers = F.normalize(source_centers, dim=1)
model.module.head.weight.data[0:n_s_classes].copy_(
source_centers.to(device))
# save memory
del source_centers, cluster_source_loader, source_features_per_id
# define optimizer and lr scheduler
optimizer = Adam(model.module.get_parameters(base_lr=args.lr,
rate=args.rate),
args.lr,
weight_decay=args.weight_decay)
lr_scheduler = StepLR(optimizer, step_size=args.step_size, gamma=0.1)
if args.resume:
checkpoint = torch.load(args.resume, map_location='cpu')
utils.copy_state_dict(model, checkpoint['model'])
optimizer.load_state_dict(checkpoint['optimizer'])
lr_scheduler.load_state_dict(checkpoint['lr_scheduler'])
args.start_epoch = checkpoint['epoch'] + 1
# start training
best_test_mAP = 0.
for epoch in range(args.start_epoch, args.epochs):
# run clustering algorithm and generate pseudo labels
train_target_iter = run_dbscan(cluster_target_loader, model,
target_dataset, train_transform, args)
# train for one epoch
print(lr_scheduler.get_lr())
train(train_source_iter, train_target_iter, model, optimizer, xbm,
epoch, args)
if (epoch + 1) % args.eval_step == 0 or (epoch == args.epochs - 1):
# remember best mAP and save checkpoint
torch.save(
{
'model': model.state_dict(),
'optimizer': optimizer.state_dict(),
'lr_scheduler': lr_scheduler.state_dict(),
'epoch': epoch
}, logger.get_checkpoint_path(epoch))
print("Test on target domain...")
_, test_mAP = validate(test_loader,
model,
target_dataset.query,
target_dataset.gallery,
device,
cmc_flag=True,
rerank=args.rerank)
if test_mAP > best_test_mAP:
shutil.copy(logger.get_checkpoint_path(epoch),
logger.get_checkpoint_path('best'))
best_test_mAP = max(test_mAP, best_test_mAP)
# update lr
lr_scheduler.step()
print("best mAP on target = {}".format(best_test_mAP))
logger.close()
def run_exp(argsdict):
# Example of usage of the code provided and recommended hyper parameters for training GANs.
data_root = './'
device = 'cuda' if torch.cuda.is_available() else 'cpu'
n_iter = 50 # N training iterations
n_critic_updates = 1 # N critic updates per generator update
train_batch_size = argsdict['batch_size']
lr = 1e-4
beta1 = 0.5
beta2 = 0.9
z_dim = 25
hidden_dim = (400, 400)
if argsdict['dataset'] in ['svhn']:
image_shape = (3, 32, 32)
encoding = 'tanh'
elif argsdict['dataset'] in ['CIFAR']:
image_shape = (3, 32, 32)
encoding = 'sigmoid'
elif argsdict['dataset'] in ['MNIST']:
image_shape = (1, 28, 28)
encoding = 'sigmoid'
elif argsdict['dataset'] in ['Gaussian']:
image_shape = (1, 28, 28)
encoding = 'tanh'
# Finding random mu
mus = []
for gaus in range(argsdict['number_gaussians']):
mus.append([random.randint(0, 27) for _ in range(argsdict['Gauss_size'])])
mus = torch.tensor(mus)
argsdict['mus'] = mus
# Use the GPU if you have one
if torch.cuda.is_available():
print("Using the GPU")
device = torch.device("cuda")
else:
print("WARNING: You are about to run on cpu, and this will likely run out \
of memory. \n You can try setting batch_size=1 to reduce memory usage")
device = torch.device("cpu")
train_loader, valid_loader, test_loader, num_samples = get_data(argsdict)
print(device)
print(num_samples)
generator = Generator(image_shape, hidden_dim[0], hidden_dim[1], z_dim, encoding).to(device)
# generator = Generatorsvhn(z_dim, hidden_dim).to(device)
critic = Critic(image_shape, 400, 400).to(device)
# critic = Criticsvhn(argsdict['hidden_discri_size']).to(device)
# TODO Adding beta seems to make total variation go to 0, why.
# TODO In rapport talk about how finicky the whole system is
optim_critic = optim.Adam(critic.parameters(), lr=lr) # , betas=(beta1, beta2))
optim_generator = optim.Adam(generator.parameters(), lr=lr) # , betas=(beta1, beta2))
losses = Divergence(argsdict['divergence'])
if argsdict['use_cuda']:
Fix_Noise = Variable(torch.normal(torch.zeros(25, z_dim), torch.ones(25, z_dim))).cuda()
else:
Fix_Noise = Variable(torch.normal(torch.zeros(25, z_dim), torch.ones(25, z_dim)))
losses_Generator = []
losses_Discriminator = []
real_statistics = []
fake_statistics = []
# COMPLETE TRAINING PROCEDURE
for epoch in range(n_iter):
G_losses, D_losses = [], []
real_stat, fake_stat = [], []
if argsdict['visualize']:
real_imgs = torch.zeros([num_samples, image_shape[1], image_shape[2]])
for i_batch, sample_batch in enumerate(train_loader):
optim_critic.zero_grad()
if argsdict['use_cuda']:
real_img, label_batch = sample_batch[0].cuda(), sample_batch[1]
else:
real_img, label_batch = sample_batch[0], sample_batch[1]
if argsdict['visualize']:
real_imgs[i_batch * train_batch_size:i_batch * train_batch_size + train_batch_size] = real_img.squeeze(
1)
# fake img
if argsdict['use_cuda']:
noise = Variable(
torch.normal(torch.zeros(train_batch_size, z_dim), torch.ones(train_batch_size, z_dim))).cuda()
else:
noise = Variable(
torch.normal(torch.zeros(train_batch_size, z_dim), torch.ones(train_batch_size, z_dim)))
fake_img = generator(noise)
# Attempting loss
DX_score = critic(real_img)
DG_score = critic(fake_img)
loss_D = losses.D_loss(DX_score, DG_score)
fake, real = losses.RealFake(DG_score, DX_score)
real_stat.append(real)
fake_stat.append(fake)
loss_D.backward()
# D_grad=critic.x[0].weight.grad.detach()
optim_critic.step()
# Clip weights of discriminator
for p in critic.parameters():
p.data.clamp_(-0.1, 0.1)
# train the generator ever n_critic iterations
D_losses.append(loss_D.item())
if i_batch % n_critic_updates == 0:
optim_generator.zero_grad()
gen_img = generator(noise)
if argsdict['modified_loss']:
DG_score = critic(gen_img)
# We maximize instead of minimizing
loss_G = losses.G_loss_modified_sec_32(DG_score)
else:
DG_score = critic(gen_img)
loss_G = losses.G_loss(DG_score)
loss_G.backward()
optim_generator.step()
G_losses.append(loss_G.item())
# print(G_losses)
# print(D_losses)
# print(D_grad)
print("Epoch[%d/%d], G Loss: %.4f, D Loss: %.4f"
% (epoch, n_iter, np.mean(G_losses), np.mean(D_losses)))
print(
f"Classified on average {round(np.mean(real_stat), 2)} real examples correctly and {round(np.mean(fake_stat), 2)} fake examples correctly")
losses_Generator.append(np.mean(G_losses))
losses_Discriminator.append(np.mean(D_losses))
real_statistics.append(np.mean(real_stat))
fake_statistics.append(np.mean(fake_stat))
if argsdict['dataset'] == 'Gaussian':
# A bit hacky but reset iterators
train_loader, valid_loader, test_loader = get_data(argsdict)
if argsdict['visualize']:
if argsdict['use_cuda']:
noise = Variable(torch.normal(torch.zeros(500, z_dim), torch.ones(500, z_dim))).cuda()
else:
noise = Variable(torch.normal(torch.zeros(500, z_dim), torch.ones(500, z_dim)))
fake_imgs = generator(noise)
visualize_tsne(fake_imgs, real_imgs[:500], argsdict, epoch)
with torch.no_grad():
img = generator(Fix_Noise)
# Saving Images
if argsdict['modified_loss']:
save_image(img.view(-1, image_shape[0], image_shape[1], image_shape[2]),
f"{argsdict['dataset']}_IMGS/{argsdict['divergence']}/GRID_trick32%d.png" % epoch, nrow=5,
normalize=True)
else:
save_image(img.view(-1, image_shape[0], image_shape[1], image_shape[2]),
f"{argsdict['dataset']}_IMGS/{argsdict['divergence']}/GRID%d.png" % epoch, nrow=5,
normalize=True)
with open(f"{argsdict['dataset']}_IMGS/{argsdict['divergence']}/Losses.txt", "w") as f:
json.dump({'Gen_Loss': losses_Generator, 'Discri_Loss': losses_Discriminator, 'real_stat': real_statistics,
'fake_stat': fake_statistics}, f)
# Update the losses plot every 5 epochs
if epoch % 5 == 0 and epoch != 0:
plot_losses(argsdict, epoch + 1, show_plot=0)
plot_losses(argsdict, n_iter)