def load_model(model):
# build model
net = get_pose_net()
if torch.cuda.is_available():
net = net.cuda()
# net = nn.DataParallel(net) # multi-Gpu
save_path = os.path.join('ckpt2/ucihand_lstm_pm' + str(model)+'.pth')
state_dict = torch.load(save_path)
net.load_state_dict(state_dict)
return net
os.mkdir(save_dir)
transform = transforms.Compose([transforms.ToTensor()])
# Build dataset
train_data = Dhp19PoseDataset(data_dir=train_data_dir,
label_dir=train_label_dir,
temporal=temporal,
train=True)
print('Train dataset total number of images sequence is ----' +
str(len(train_data)))
# Data Loader
train_dataset = DataLoader(train_data, batch_size=batch_size, shuffle=False)
net = get_pose_net()
#gpus = [int(i) for i in '2'.split(',')]
model = torch.nn.DataParallel(net, device_ids=[0]).cuda()
def train():
criterion = JointsMSELoss(
use_target_weight=config.LOSS.USE_TARGET_WEIGHT).cuda()
optimizer = get_optimizer(config, model)
lr_scheduler = torch.optim.lr_scheduler.MultiStepLR(
optimizer, config.TRAIN.LR_STEP, config.TRAIN.LR_FACTOR)
for epoch in range(begin_epoch, epochs + 1):
batch_time = AverageMeter()
data_time = AverageMeter()
losses = AverageMeter()
acc = AverageMeter()
# vgg11 = models.vgg11().cuda() # summary(vgg11,(3,224,224)) # vgg16 = models.vgg16().cuda() # summary(vgg16,(3,224,224)) # vgg19 = models.vgg19().cuda() # summary(vgg19,(3,224,224)) # alexnet = models.alexnet().cuda() # summary(alexnet,(3,224,224)) # resnet18 = models.resnet18().cuda() # summary(resnet18,(3,224,224)) # resnet34= models.resnet34().cuda() # summary(resnet34,(3,224,224)) # from torch import nn # lstm = nn.LSTM(4,10).cuda() # print(lstm.) # nn. # net = get_pose_net().cuda() # summary(net,(1, 368, 368)) # print(net) import torch net = get_pose_net().cuda() # a = torch.randn(1, 3, 224, 224).cuda() # batch size = 2 a = torch.randn(1, 1, 224, 224).cuda() # batch size = 2 net(a)