def init_dataloaders_1s_mt(args,
batch_size,
num_workers,
is_bgr=True,
is_shuffle=True):
# source dataset
imdb, roidb, ratio_list, ratio_index = combined_roidb(args.imdb_name)
train_size = len(roidb)
sampler_batch = sampler(train_size, batch_size)
dataset_s = roibatchLoader(roidb, ratio_list, ratio_index, batch_size, \
imdb.num_classes, training=True, is_bgr=is_bgr)
dataloader_s = torch.utils.data.DataLoader(dataset_s,
batch_size=batch_size,
sampler=sampler_batch,
num_workers=num_workers)
m_dataloader_t = []
m_imdb_t = []
# target dataset
for i_n_t in args.imdb_name_target:
imdb_t, roidb_t, ratio_list_t, ratio_index_t = combined_roidb(i_n_t)
m_imdb_t.append(imdb_t)
train_size_t = len(roidb_t)
sampler_batch_t = sampler(train_size_t, batch_size)
dataset_t = roibatchLoader(roidb_t, ratio_list_t, ratio_index_t, batch_size, \
imdb.num_classes, training=True, is_bgr=is_bgr)
dataloader_t = torch.utils.data.DataLoader(dataset_t,
batch_size=batch_size,
sampler=sampler_batch_t,
num_workers=num_workers)
m_dataloader_t.append(dataloader_t)
return dataloader_s, m_dataloader_t, imdb, m_imdb_t
def init_dataloaders_1s_1t(args,
batch_size,
num_workers,
is_bgr=True,
is_train=True):
# source dataset
imdb, roidb, ratio_list, ratio_index = combined_roidb(args.imdb_name)
train_size = len(roidb)
sampler_batch = sampler(train_size, batch_size)
dataset_s = roibatchLoader(roidb, ratio_list, ratio_index, batch_size, \
imdb.num_classes, training=is_train, is_bgr=is_bgr)
dataloader_s = torch.utils.data.DataLoader(dataset_s,
batch_size=batch_size,
sampler=sampler_batch,
num_workers=num_workers)
# target dataset
dataloader_t = None
imdb_t = None
if args.dataset_t != "":
imdb_t, roidb_t, ratio_list_t, ratio_index_t = combined_roidb(
args.imdb_name_target)
train_size_t = len(roidb_t)
sampler_batch_t = sampler(train_size_t, batch_size)
dataset_t = roibatchLoader(roidb_t, ratio_list_t, ratio_index_t, batch_size, \
imdb.num_classes, training=is_train, is_bgr=is_bgr)
dataloader_t = torch.utils.data.DataLoader(dataset_t,
batch_size=batch_size,
sampler=sampler_batch_t,
num_workers=num_workers)
return dataloader_s, dataloader_t, imdb, imdb_t
def init_dataloaders_1s_mixed_mt(args, batch_size, num_workers):
# source dataset
imdb, roidb, ratio_list, ratio_index = combined_roidb(args.imdb_name)
train_size = len(roidb)
sampler_batch = sampler(train_size, batch_size)
dataset_s = roibatchLoader(roidb, ratio_list, ratio_index, batch_size, \
imdb.num_classes, training=True)
dataloader_s = torch.utils.data.DataLoader(dataset_s,
batch_size=batch_size,
sampler=sampler_batch,
num_workers=num_workers)
m_datasets = []
m_imdb_t = []
total_train_size = 0
# target dataset
for i_n_t in args.imdb_name_target:
imdb_t, roidb_t, ratio_list_t, ratio_index_t = combined_roidb(i_n_t)
m_imdb_t.append(imdb_t)
total_train_size += len(roidb_t)
dataset_t = roibatchLoader(roidb_t, ratio_list_t, ratio_index_t, batch_size, \
imdb.num_classes, training=True)
m_datasets.append(dataset_t)
sampler_batch_t = sampler(total_train_size, batch_size)
concat_dataset = torch.utils.data.ConcatDataset(m_datasets)
dataloader_t = torch.utils.data.DataLoader(concat_dataset,
batch_size=batch_size,
sampler=sampler_batch_t,
num_workers=num_workers)
return dataloader_s, dataloader_t, imdb
# source dataset
imdb, roidb, ratio_list, ratio_index = combined_roidb(args.imdb_name)
train_size = len(roidb)
# target dataset
imdb_t, roidb_t, ratio_list_t, ratio_index_t = combined_roidb(
args.imdb_name_target)
train_size_t = len(roidb_t)
print('{:d} source roidb entries'.format(len(roidb)))
print('{:d} target roidb entries'.format(len(roidb_t)))
output_dir = args.save_dir + "/" + args.net + "/" + args.dataset
if not os.path.exists(output_dir):
os.makedirs(output_dir)
sampler_batch = sampler(train_size, args.batch_size)
sampler_batch_t = sampler(train_size_t, args.batch_size)
dataset_s = roibatchLoader(roidb, ratio_list, ratio_index, args.batch_size, \
imdb.num_classes, training=True)
dataloader_s = torch.utils.data.DataLoader(dataset_s,
batch_size=args.batch_size,
sampler=sampler_batch,
num_workers=args.num_workers)
dataset_t = roibatchLoader(roidb_t, ratio_list_t, ratio_index_t, args.batch_size, \
imdb.num_classes, training=True)
dataloader_t = torch.utils.data.DataLoader(dataset_t,
batch_size=args.batch_size,
sampler=sampler_batch_t,
num_workers=args.num_workers)
