def trainCNN(modelName='resnet'):
# More hyperparameters
dataset = ImageDataset()
dataset_labels = dataset.get_all_labels()
num_classes = len(dataset_labels)
if modelName == 'resnet':
model = resnet_dropout_18(num_classes=num_classes, p=cnnDropout)
elif modelName == 'inception':
model = Inception3(num_classes=num_classes, aux_logits=False)
elif modelName == 'segnet':
# TODO: Figure out how dims need to be changed based off of NYU dataset
model = SegNet(input_channels=3,
output_channels=1,
pretrained_vgg=True)
else:
raise Exception("Please select one of \'resnet\' or \'inception\' or "
"\'segnet\'")
if torch.cuda.is_available():
if multiGPU:
model = nn.DataParallel(model)
model = model.cuda()
optimizer = optim.Adam(model.parameters(), lr=cnnLr)
scheduler = torch.optim.lr_scheduler.ReduceLROnPlateau(optimizer,
'min',
patience=2)
# setup the device for running
device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
model = model.to(device)
model.eval()
return model
# test_loader = DataLoader(dataset=cv_dataset, batch_size=10, shuffle=True, num_workers=2)
# cv_dataset = CamVidDataset('./CamVid/', mold='test', transforms=transformations, output_size=(352, 480), predct=True)
cv_dataset = DatasetFrombaidu('./baidu/',
mold='val',
transforms=transformations,
output_size=(720, 720),
predct=True)
# cv_loader = DataLoader(dataset=cv_dataset, batch_size=1, shuffle=False, num_workers=8, drop_last=True)
# 定义预测函数
# print(cv_dataset.datalen)
# cm = np.array(COLORMAP).astype('uint8')
# cm = np.array(CamVid_colours).astype('uint8')
n_class = 9
net = SegNet(num_classes=9)
# net=SegNet(num_classes=12)
net.cuda()
net.eval()
dir = './checkpoints/baiduSegNet5.pth'
state = t.load(dir)
net.load_state_dict(state['net'])
test_data, test_label = cv_dataset[1]
print(test_data.size())
# out=net(Variable(test_data.unsqueeze(0)).cuda())
# print(out.data.size())
# pred = out.max(1)[1].squeeze().cpu().data.numpy()
# print(pred.shape)
def predict(im, label): # 预测结果
im = Variable(im.unsqueeze(0)).cuda()
for i, key in enumerate(class_encoding.keys()):
print("{} \t {}".format(i, key))
optimizer = optim.SGD(own_net.parameters(),
lr=1e-3,
weight_decay=5e-4,
momentum=0.9)
# Evaluation metric
ignore_index = list(class_encoding).index('unlabeled')
criterion = nn.CrossEntropyLoss(ignore_index=ignore_index)
if use_gpu:
own_net = own_net.cuda()
trainer = train.Trainer(criterion,
optimizer,
own_net,
single_sample=single_sample)
chkpt_path = "D:\CloudStorage\ODWork\Work Dropbox\MB\Tmp\SGD_lr_0.001_momentum_0.9_dampening_0_weight_decay_0.0005_nesterov_False_CrossEntropyLoss_RGB_2\Checkpoints\epoch150.ckpt"
trainer.load_checkpoint(chkpt_path)
own_net = trainer.model
own_net.eval()
for i, (inputs, labels) in enumerate(trainloader, 0):