pkl.HIGHEST_PROTOCOL)
own_state = net.state_dict()
for name, param in pret_net.items():
if name not in own_state:
continue
if isinstance(param, Parameter):
param = param.data
try:
own_state[name].copy_(param)
print('Copied {}'.format(name))
except:
print('Did not find {}'.format(name))
continue
# Move model to GPU and set train mode
net.load_state_dict(own_state)
net.cuda()
net.train()
# TODO: Create optimizer for network parameters from conv2 onwards
# (do not optimize conv1)
if not os.path.exists(output_dir):
os.makedirs(output_dir)
# training
train_loss = 0
tp, tf, fg, bg = 0., 0., 0, 0
step_cnt = 0
re_cnt = False
t = Timer()
for name, param in pret_net.items():
if name not in own_state:
continue
if isinstance(param, Parameter):
param = param.data
try:
own_state[name].copy_(param)
print('Copied {}'.format(name))
except:
print('Did not find {}'.format(name))
continue
resume_file = "./wsddn_test_checkpoint_adam"
if resume and os.path.isfile(resume_file):
print("=> loading checkpoint '{}'".format(resume_file))
checkpoint = torch.load(resume_file)
epoch = checkpoint['epoch']
net.load_state_dict(checkpoint['state_dict'])
print("=> loaded checkpoint '{}' (epoch {})"
.format(resume_file, checkpoint['epoch']))
# Move model to GPU and set train mode
net.cuda()
net.eval()
save_name = '{}_{}'
aps = test_net(save_name, net, test_imdb,
max_per_image, thresh=thresh, visualize=visualize,logger = logger, step = epoch)
print("here's the aps")
print(aps)