def main():
global arg
arg = parser.parse_args()
print(arg)
#Prepare DataLoader
data_loader = dataloader.Motion_DataLoader(
BATCH_SIZE=arg.batch_size,
num_workers=0,
path='../bold_data/BOLD_ijcv/BOLD_public/optic_flow/',
ucf_list = '../bold_data/BOLD_ijcv/BOLD_public/annotations/',
ucf_split = '04',
in_channel=10,
)
train_loader,test_loader, test_video = data_loader.run()
#Model
model = Motion_CNN(
# Data Loader
train_loader=train_loader,
test_loader=test_loader,
# Utility
start_epoch=arg.start_epoch,
resume=arg.resume,
evaluate=arg.evaluate,
# Hyper-parameter
nb_epochs=arg.epochs,
lr=arg.lr,
batch_size=arg.batch_size,
channel = 10,
test_video=test_video
)
#Training
model.run()
def main():
global arg
arg = parser.parse_args()
print(arg)
#Prepare DataLoader
data_loader = dataloader.Motion_DataLoader(
BATCH_SIZE=arg.batch_size,
num_workers=8,
path=r"/mnt/disks/datastorage/videos/flownet2/",
ucf_list=r"/home/mlp/two-stream-action-recognition/UCF_list/",
ucf_split='01',
in_channel=10,
)
train_loader, test_loader, test_video = data_loader.run()
#Model
model = Motion_CNN(
# Data Loader
train_loader=train_loader,
test_loader=test_loader,
# Utility
start_epoch=arg.start_epoch,
resume=arg.resume,
evaluate=arg.evaluate,
# Hyper-parameter
nb_epochs=arg.epochs,
lr=arg.lr,
batch_size=arg.batch_size,
channel=10 * 2,
test_video=test_video,
nb_classes=arg.nb_classes, # added the nb_classes
finetune=arg.finetune)
#Training
model.run()
def main():
global arg
arg = parser.parse_args()
print arg
#Prepare DataLoader
data_loader = dataloader.Motion_DataLoader(
BATCH_SIZE=arg.batch_size,
num_workers=8, #use number of work 8
path=
'/media/semanticslab11/hdd1/data/two-stream-action/data/tvl1_flow/',
ucf_list='UCF_list/',
ucf_split='01',
in_channel=10,
bidirectional=arg.bidirectional)
train_loader, test_loader, test_video = data_loader.run()
#Model
model = Motion_CNN(
# Data Loader
train_loader=train_loader,
test_loader=test_loader,
# Utility
start_epoch=arg.start_epoch,
resume=arg.resume,
evaluate=arg.evaluate,
# Hyper-parameter
nb_epochs=arg.epochs,
lr=arg.lr,
batch_size=arg.batch_size,
channel=10 * 2,
test_video=test_video)
#Training
model.run()
def main():
global arg
arg = parser.parse_args()
print(arg)
#Prepare DataLoader
data_loader = dataloader.Motion_DataLoader(
BATCH_SIZE=arg.batch_size,
num_workers=8,
path='/home/zdadadaz/Desktop/course/medical/data/UCF101_flow/tvl1_flow',
ucf_list='./UCF_list/',
ucf_split='01',
in_channel=10,
root_path='../')
train_loader, test_loader, test_video = data_loader.run()
#Model
model = Motion_CNN(
# Data Loader
train_loader=train_loader,
test_loader=test_loader,
# Utility
start_epoch=arg.start_epoch,
resume=arg.resume,
evaluate=arg.evaluate,
# Hyper-parameter
nb_epochs=arg.epochs,
lr=arg.lr,
batch_size=arg.batch_size,
channel=10 * 2,
test_video=test_video)
#Training
model.run()
def main():
global arg
arg = parser.parse_args()
print arg
#Prepare DataLoader
data_loader = dataloader.Motion_DataLoader(
BATCH_SIZE=arg.batch_size,
num_workers=8,
path='/home/lili/Video/datasets/HMDB51_concise/',
train_ucf_list='/home/lili/Video/datasets/HMDB51_concise/list/',
test_ucf_list=
'/ubc/cs/research/tracking-raid/candice/datasets/HMDB51/list/new_test_flow.list',
ucf_split='01',
in_channel=10,
)
train_loader, test_loader, test_video = data_loader.run()
#Model
model = Motion_CNN(
# Data Loader
train_loader=train_loader,
test_loader=test_loader,
# Utility
start_epoch=arg.start_epoch,
resume=arg.resume,
evaluate=arg.evaluate,
# Hyper-parameter
nb_epochs=arg.epochs,
lr=arg.lr,
batch_size=arg.batch_size,
channel=10 * 2,
test_video=test_video)
#Training
model.run()
def main():
global arg
arg = parser.parse_args()
print arg
#Prepare DataLoader
data_loader = dataloader.Motion_DataLoader(
BATCH_SIZE=arg.batch_size,
num_workers=8,
path='/home/ubuntu/data/UCF101/tvl1_flow/',
ucf_list=
'/home/ubuntu/cvlab/pytorch/ucf101_two_stream/github/UCF_list/',
ucf_split='01',
in_channel=10,
)
train_loader, test_loader, test_video = data_loader.run()
#Model
model = Motion_CNN(
# Data Loader
train_loader=train_loader,
test_loader=test_loader,
# Utility
start_epoch=arg.start_epoch,
resume=arg.resume,
evaluate=arg.evaluate,
# Hyper-parameter
nb_epochs=arg.epochs,
lr=arg.lr,
batch_size=arg.batch_size,
channel=10 * 2,
test_video=test_video)
#Training
model.run()
def main():
global arg
arg = parser.parse_args()
print arg
#Prepare DataLoader
data_loader = dataloader.Motion_DataLoader(
BATCH_SIZE=arg.batch_size,
num_workers=8,
path='/data/MM1/aps1/aniru/aniru/repo/dataset/UCF101/tvl1_flow',
ucf_list='//data/MM1/aps1/aniru/aniru/repo/actionRecognition/UCF_list',
ucf_split='01',
in_channel=10,
)
train_loader, test_loader, test_video = data_loader.run()
#Model
model = Motion_CNN(
# Data Loader
train_loader=train_loader,
test_loader=test_loader,
# Utility
start_epoch=arg.start_epoch,
resume=arg.resume,
evaluate=arg.evaluate,
# Hyper-parameter
nb_epochs=arg.epochs,
lr=arg.lr,
batch_size=arg.batch_size,
channel=10 * 2,
test_video=test_video)
#Training
model.run()
def main():
global arg
arg = parser.parse_args()
print(arg)
#Prepare DataLoader
data_loader = dataloader.Motion_DataLoader(
BATCH_SIZE=arg.batch_size,
num_workers=8,
path='/data/tvl1_flow/',
ucf_list=
'/media/lsc/DATA/github/two-stream-action-recognition/UCF_list/',
ucf_split='01',
in_channel=10,
)
train_loader, test_loader, test_video = data_loader.run()
#Model
model = Motion_CNN(
# Data Loader
train_loader=train_loader,
test_loader=test_loader,
# Utility
start_epoch=arg.start_epoch,
resume=arg.resume,
evaluate=arg.evaluate,
# Hyper-parameter
nb_epochs=arg.epochs,
lr=arg.lr,
batch_size=arg.batch_size,
channel=10 * 2,
test_video=test_video)
#Training
model.run()
def main():
global arg
arg = parser.parse_args()
print arg
# Prepare DataLoader
data_loader = dataloader.Motion_DataLoader(
BATCH_SIZE=arg.batch_size,
num_workers=8,
path='/hdd/UCF-101/Data/optical-flow/',
ucf_list=os.getcwd() + '/UCF_data_references/',
ucf_split='01',
in_channel=10,
)
train_loader, test_loader, test_video = data_loader.run()
# Model
model = Motion_CNN(
# Data Loader
train_loader=train_loader,
test_loader=test_loader,
# Utility
start_epoch=arg.start_epoch,
resume=arg.resume,
evaluate=arg.evaluate,
# Hyper-parameter
nb_epochs=arg.epochs,
lr=arg.lr,
batch_size=arg.batch_size,
channel=10 * 2,
test_video=test_video,
demo=arg.demo)
# Training
model.run()
def main():
global arg
arg = parser.parse_args()
print(arg)
#Prepare DataLoader
data_loader = dataloader.Motion_DataLoader(
BATCH_SIZE=arg.batch_size,
num_workers=8,
path='/home/yzy20161103/csce636_project/project/opt_475/',
ucf_list='/home/yzy20161103/csce636_project/project/UCF_list/',
ucf_split='01',
in_channel=10,
)
train_loader, test_loader, test_video = data_loader.run()
#Model
model = Motion_CNN(
# Data Loader
train_loader=train_loader,
test_loader=test_loader,
# Utility
start_epoch=arg.start_epoch,
resume=arg.resume,
evaluate=arg.evaluate,
# Hyper-parameter
nb_epochs=arg.epochs,
lr=arg.lr,
batch_size=arg.batch_size,
channel=10 * 2,
test_video=test_video)
#Training
model.run()
def main():
global arg
arg = parser.parse_args()
print(arg)
#Prepare DataLoader
data_loader = dataloader.Motion_DataLoader(BATCH_SIZE=arg.batch_size,
num_workers=8,
path='/home/hosseing/datasets/tvl1_flow/',
ucf_list='/home/hosseing/REPOS/two stream/two-stream-action-recognition//UCF_list/',
ucf_split='01',
in_channel=10,)
train_loader,test_loader, test_video = data_loader.run()
# Building model
model = Motion_CNN(
# Data Loader
train_loader=train_loader,
test_loader=test_loader,
# Utility
start_epoch=arg.start_epoch,
resume=arg.resume,
evaluate=arg.evaluate,
# Hyper-parameter
nb_epochs=arg.epochs,
lr=arg.lr,
batch_size=arg.batch_size,
channel = 10*2,
test_video=test_video
)
# Training
with torch.cuda.device(0):
model.run()
def main():
global arg
arg = parser.parse_args()
print arg
#Prepare DataLoader
data_loader = dataloader.Motion_DataLoader(
BATCH_SIZE=arg.batch_size,
num_workers=12,
path='/home/ubuntu/data/UCF101_optical_flow/step1/',
ucf_list='/home/bassel/data/ucfTrainTestlist/',
ucf_split='01',
in_channel=10,
)
train_loader,test_loader, test_video = data_loader.run()
#Model
model = Motion_CNN(
# Data Loader
train_loader=train_loader,
test_loader=test_loader,
# Utility
start_epoch=arg.start_epoch,
resume=arg.resume,
evaluate=arg.evaluate,
# Hyper-parameter
nb_epochs=arg.epochs,
lr=arg.lr,
batch_size=arg.batch_size,
channel = 10*2,
test_video=test_video
)
#Training
model.run()
def main():
global arg
arg = parser.parse_args()
print(arg)
#Prepare DataLoader
data_loader = dataloader.Motion_DataLoader(
BATCH_SIZE=arg.batch_size,
num_workers=0,
path='D:\\Data\\UCF101\\ucf101_tvl1_flow\\tvl1_flow\\',
ucf_list=
'D:\\Radha\\Downloads\\two-stream-action-recognition-master\\UCF_list\\',
ucf_split='01',
in_channel=10,
)
train_loader, test_loader, test_video = data_loader.run()
#Model
model = Motion_CNN(
# Data Loader
train_loader=train_loader,
test_loader=test_loader,
# Utility
start_epoch=arg.start_epoch,
resume=arg.resume,
evaluate=arg.evaluate,
# Hyper-parameter
nb_epochs=arg.epochs,
lr=arg.lr,
batch_size=arg.batch_size,
channel=10 * 2,
test_video=test_video)
#Training
model.run()
def main():
global arg
arg = parser.parse_args()
print arg
#Prepare DataLoader
data_loader = dataloader.Motion_DataLoader(
BATCH_SIZE=arg.batch_size,
num_workers=4,
path='/home/niloofar/Work/Data/Motion/UCF-101/',
ucf_list=os.getcwd()+'/UCF_list/',
ucf_split='01',
in_channel=10,
)
data_loader_my = dataloader.Motion_DataLoader(
BATCH_SIZE=arg.batch_size_my,
num_workers=4,
path='/home/niloofar/Work/Data/Motion/UCF-101/',
ucf_list=os.getcwd()+'/My_list/',
ucf_split='01',
in_channel=10,
)
train_loader,test_loader, test_video = data_loader.run()
train_loader_my, test_loader_my, test_video_my = data_loader_my.run()
#Model
model = Motion_CNN(
nb_epochs=arg.epochs,
lr=arg.lr,
batch_size=arg.batch_size,
batch_size_my=arg.batch_size_my,
resume=arg.resume,
start_epoch=arg.start_epoch,
evaluate=arg.evaluate,
demo=arg.demo,
channel = 10*2,
train_loader=train_loader,
train_loader_my=train_loader_my,
test_loader=test_loader,
test_loader_my=test_loader_my,
test_video=test_video,
test_video_my=test_video_my
)
#Training
model.run()
def main():
global arg
args = parser.parse_args()
print(args)
#Prepare DataLoader
train_loader_obj = dataloader.Motion_DataLoader(
is_val=False,
batch_size=args.batch_size,
num_workers=1,
root_dir=args.data_root,
desc_numpy_file=args.train_desc,
in_channel=SEQ_FRAME_QTY)
val_loader_obj = dataloader.Motion_DataLoader(
is_val=True,
batch_size=args.batch_size,
num_workers=1,
root_dir=args.data_root,
desc_numpy_file=args.test_desc,
in_channel=SEQ_FRAME_QTY)
train_loader = train_loader_obj.get_loader()
val_loader = val_loader_obj.get_loader()
model = Motion_CNN(
# Data Loader
train_loader=train_loader,
test_loader=val_loader,
# Utility
start_epoch=args.start_epoch,
resume=args.resume,
evaluate=args.evaluate,
# Hyper-parameter
nb_epochs=args.epochs,
lr=args.lr,
patience=args.patience,
batch_size=args.batch_size,
nb_classes=args.class_qty,
nb_classes_new=args.class_qty_new,
channel=SEQ_FRAME_QTY * 2)
#Training
model.run()
def main():
global arg
arg = parser.parse_args()
print arg
if arg.checkpoint_path:
if not os.path.exists(arg.checkpoint_path):
os.makedirs(arg.checkpoint_path)
if arg.weight_per_class:
weights = np.array(arg.weight_per_class)
weights = torch.FloatTensor(weights)
#Prepare DataLoader
data_loader = dataloader.Motion_DataLoader(
BATCH_SIZE=arg.batch_size,
num_workers=8,
path=
'/home/candice/Documents/dataset_icehockey/events/flipped_data_correction/flow',
train_ucf_list=
'/home/candice/Documents/dataset_icehockey/events/experiments/new/train_list.txt',
test_ucf_list=
'/home/candice/Documents/dataset_icehockey/events/experiments/new/test_list.txt',
ucf_split='01',
in_channel=arg.num_imgs,
frames_list=arg.frames_list)
train_loader, test_loader, test_video = data_loader.run()
#Model
model = Motion_CNN(
# Data Loader
train_loader=train_loader,
test_loader=test_loader,
# Utility
start_epoch=arg.start_epoch,
resume=arg.resume,
evaluate=arg.evaluate,
# Hyper-parameter
nb_epochs=arg.epochs,
lr=arg.lr,
batch_size=arg.batch_size,
channel=2 * arg.num_imgs,
test_video=test_video,
weights=weights)
#Training
model.run()
def main():
global arg
arg = parser.parse_args()
print(arg)
#Prepare DataLoader
data_loader = dataloader.Motion_DataLoader(
BATCH_SIZE=arg.batch_size,
num_workers=8,
path='/research/dept2/mli/Data/tvl1_flow/',
ucf_list='./UCF_list/',
ucf_split='01',
in_channel=10,
)
train_loader, test_loader, test_video = data_loader.run()
#Model
model = Motion_CNN(
# Data Loader
train_loader=train_loader,
test_loader=test_loader,
# Utility
start_epoch=arg.start_epoch,
resume=arg.resume,
evaluate=arg.evaluate,
# Hyper-parameter
nb_epochs=arg.epochs,
lr=arg.lr,
batch_size=arg.batch_size,
channel=10 * 2,
test_video=test_video)
#Training
if arg.evaluate:
cudnn.benchmark = True
model.build_model()
model.resume_and_evaluate()
else:
model.run()
from utils import *
import torch.optim as optim
import dataloader
import torch
import os
from ResidualNetwork import *
os.environ["CUDA_VISIBLE_DEVICES"] = "0"
# Prepare DataLoader
print '*************************Prepare DataLoader********************************'
data_loader = dataloader.Motion_DataLoader(
BATCH_SIZE=10,
num_workers=8,
path='/media/ming/DATADRIVE1/UCF101 Dataset/tvl1_flow/',
UCF_list='/media/ming/DATADRIVE1/UCF101 Multi-stream Code/UCF_list/',
in_channel=10 # NEW ADD
)
train_loader, test_loader, test_video = data_loader.run()
# build the model
model = resnet50(pretrained=False, channel=10 * 2).cuda()
criterion = nn.CrossEntropyLoss().cuda()
optimizer = optim.SGD(model.parameters(), lr=0.01, momentum=0.9)
def train_1epoch(epoch):
print '---------train execute ', epoch, ' epoch'
