def __init__(self,
annFile: str,
imageDir: str,
targetHeight: int,
targetWidth: int,
numClass: int,
train: bool = True):
self.annotations = {}
self.table = {
'不良-機械傷害': 0,
'不良-著色不佳': 1,
'不良-炭疽病': 2,
'不良-乳汁吸附': 3,
'不良-黑斑病': 4
}
self.imageDir = imageDir
self.numClass = numClass
with open(annFile, 'r', encoding='utf-8-sig') as f:
for line in f.readlines():
arr = line.rstrip().split(',')
ans = []
for idx in range(1, len(arr), 5):
tlx, tly, w, h, c = arr[idx:idx + 5]
if tlx:
tlx, tly, w, h = list(map(float, (tlx, tly, w, h)))
if c not in self.table:
self.table[c] = len(self.table)
cx = tlx + w / 2
cy = tly + h / 2
c = self.table[c]
ans.append(list(map(int, (cx, cy, w, h, c))))
self.annotations[arr[0]] = ans
self.names = list(self.annotations)
with open('table.txt', 'w') as f:
f.write(str(self.table))
print(self.table)
if train:
self.transforms = T.Compose([
T.RandomOrder([
T.RandomHorizontalFlip(),
T.RandomVerticalFlip(),
T.RandomSizeCrop(numClass)
]),
T.Resize((targetHeight, targetWidth)),
T.ColorJitter(brightness=.2, contrast=0, saturation=0, hue=0),
T.Normalize()
])
else:
self.transforms = T.Compose(
[T.Resize((targetHeight, targetWidth)),
T.Normalize()])
def load_dataloader(opt):
if opt.imglist_file is not None:
dataset = ImageList(opt.imglist_file, transform= \
transforms.Compose([DEFAULT_TRANSFORMS, Resize(opt.img_size)]))
else:
dataset = ImageFolder(opt.image_folder, transform= \
transforms.Compose([DEFAULT_TRANSFORMS, Resize(opt.img_size)]))
dataloader = DataLoader(dataset,
batch_size=opt.batch_size,
shuffle=False,
num_workers=opt.n_cpu,
)
return dataloader
def __init__(self, root: str, annotation: str, targetHeight: int,
targetWidth: int, numClass: int):
self.root = root
self.coco = COCO(annotation)
self.ids = list(self.coco.imgs.keys())
self.targetHeight = targetHeight
self.targetWidth = targetWidth
self.numClass = numClass
self.transforms = T.Compose([
T.RandomOrder(
[T.RandomHorizontalFlip(),
T.RandomSizeCrop(numClass)]),
T.Resize((targetHeight, targetWidth)),
T.ColorJitter(brightness=.2, contrast=.1, saturation=.1, hue=0),
T.Normalize()
])
self.newIndex = {}
classes = []
for i, (k, v) in enumerate(self.coco.cats.items()):
self.newIndex[k] = i
classes.append(v['name'])
with open('classes.txt', 'w') as f:
f.write(str(classes))
def main():
val_transform_det = trans.Compose([
trans.Scale(256, 256),
])
val_data = dates.Dataset('/dataset',
'/dataset',
'/dataset/test.txt',
'val',
transform=True,
transform_med=val_transform_det)
val_loader = Data.DataLoader(val_data,
batch_size=1,
shuffle=False,
num_workers=4,
pin_memory=True)
import model.siameseNet.dares as models
model = models.SiameseNet(norm_flag='l2')
checkpoint = torch.load('the path to best model', map_location='cpu')
model.load_state_dict(checkpoint['state_dict'])
print('load success')
model = model.cuda()
save_change_map_dir = 'the path to changemap'
save_roc_dir = 'the path to roc'
time_start = time.time()
current_metric = validate(model, val_loader, save_change_map_dir,
save_roc_dir)
elapsed = round(time.time() - time_start)
elapsed = str(datetime.timedelta(seconds=elapsed))
print('Elapsed {}'.format(elapsed))
def load_data(args):
normalize = t.Normalize(mean=[0.445, 0.287, 0.190],
std=[0.31, 0.225, 0.168])
im_transform = t.Compose([t.ToTensor(), normalize])
# Use the following code fo co_transformations e.g. random rotation or random flip etc.
# co_transformer = cot.Compose([cot.RandomRotate(45)])
dsetTrain = GIANA(args.imgdir,
args.gtdir,
input_size=(args.input_width, args.input_height),
train=True,
transform=im_transform,
co_transform=None,
target_transform=t.ToLabel())
train_data_loader = data.DataLoader(dsetTrain,
batch_size=args.batch_size,
shuffle=True,
num_workers=args.num_workers)
dsetVal = GIANA(args.imgdir,
args.gtdir,
train=False,
transform=im_transform,
co_transform=None,
target_transform=t.ToLabel())
val_data_loader = data.DataLoader(dsetVal,
batch_size=args.batch_size,
shuffle=False,
num_workers=args.num_workers)
return train_data_loader, val_data_loader
def __init__(self, cms, margin_size, margin_time):
self.IMAGE_SCALE = 20
self.IMAGE_SIZE = self.IMAGE_SCALE * self.IMAGE_SCALE
self.train_transform = transforms.Compose([
transforms.Resize((self.IMAGE_SCALE, self.IMAGE_SCALE)),
transforms.ToTensor()
])
self.test_transform = transforms.Compose([
transforms.Resize((self.IMAGE_SCALE, self.IMAGE_SCALE)),
transforms.ToTensor()
])
self.CMS = cms
self.MARGIN_SIZE = margin_size
self.MARGIN_TIME = margin_time
def __init__(self,
image_dir,
crop_size=64,
blend_mode="linear",
horizontal_flip=False):
"""
:param image_dir (str): path of the images
:param crop_size (int or tuple): crop size, default is 64
:param blend_mode (str): pick on of two, `screen` or `linear`, represents image composition type
:param horizontal_flip (bool): Whether use horizontal flipping or not
"""
# super(DataLoader, self).__init__()
# 1. initialize file path or a list of file names.
assert blend_mode in ["screen", "linear"]
self.blend_mode = blend_mode
self.data_path = image_dir
self.all_filenames = os.listdir(self.data_path)
self.label_filenames = list(
filter(lambda filename: filename.startswith("norain"),
self.all_filenames))
self.num_files = len(self.label_filenames)
print("[DataLoader] preprocess {} files on dir `{}`".format(
self.num_files, self.data_path))
self.transform = transforms.Compose([
transforms.FiveCrop(
crop_size, horizontal_flip), # tuple (tl, tr, bl, br, center)
lambda crops: np.stack(
[transforms.ToArray()(crop) for crop in crops])
])
def loading_data(root, mode, batch_size=1):
mean_std = ([0.5, 0.5, 0.5], [0.25, 0.25, 0.25])
log_para = 1
if mode == 'train':
main_transform = own_transforms.Compose(
[own_transforms.RandomHorizontallyFlip()])
else:
main_transform = None
img_transform = standard_transforms.Compose([
standard_transforms.ToTensor(),
standard_transforms.Normalize(*mean_std)
])
gt_transform = standard_transforms.Compose(
[own_transforms.LabelNormalize(log_para)])
restore_transform = standard_transforms.Compose([
own_transforms.DeNormalize(*mean_std),
standard_transforms.ToPILImage()
])
dataset = HeadCountDataset(root, mode, main_transform, img_transform,
gt_transform)
if mode == 'train':
dataloader = DataLoader(dataset,
batch_size=batch_size,
shuffle=True,
drop_last=True)
else:
dataloader = DataLoader(dataset, batch_size=1, shuffle=False)
return dataloader, restore_transform
def get_transform():
transforms = []
transforms.append(custom_T.ToTensor())
transforms.append(custom_T.FasterRCNNResizer())
return custom_T.Compose(transforms)
def detect_image(model, image, img_size=416, conf_thres=0.5, nms_thres=0.5):
"""Inferences one image with model.
:param model: Model for inference
:type model: models.Darknet
:param image: Image to inference
:type image: nd.array
:param img_size: Size of each image dimension for yolo, defaults to 416
:type img_size: int, optional
:param conf_thres: Object confidence threshold, defaults to 0.5
:type conf_thres: float, optional
:param nms_thres: IOU threshold for non-maximum suppression, defaults to 0.5
:type nms_thres: float, optional
:return: Detections on image
:rtype: nd.array
"""
model.eval() # Set model to evaluation mode
# Configure input
input_img = transforms.Compose([DEFAULT_TRANSFORMS,
Resize(img_size)])((image, np.zeros(
(1, 5))))[0].unsqueeze(0)
if torch.cuda.is_available():
input_img = input_img.to("cuda")
# Get detections
with torch.no_grad():
detections = model(input_img)
detections = non_max_suppression(detections, conf_thres, nms_thres)
detections = rescale_boxes(detections[0], img_size, image.shape[:2])
return to_cpu(detections).numpy()
def get_coco(root, image_set, transforms, mode='instances'):
anno_file_template = "{}_{}2017.json"
PATHS = {
"train": ("train2017",
os.path.join("annotations",
anno_file_template.format(mode, "train"))),
"val": ("val2017",
os.path.join("annotations",
anno_file_template.format(mode, "val"))),
# "train": ("val2017", os.path.join("annotations", anno_file_template.format(mode, "val")))
}
t = [ConvertCocoPolysToMask()]
if transforms is not None:
t.append(transforms)
transforms = T.Compose(t)
img_folder, ann_file = PATHS[image_set]
img_folder = os.path.join(root, img_folder)
ann_file = os.path.join(root, ann_file)
dataset = CocoDetection(img_folder, ann_file, transforms=transforms)
if image_set == "train":
dataset = _coco_remove_images_without_annotations(dataset)
# dataset = torch.utils.data.Subset(dataset, [i for i in range(500)])
return dataset
def main():
###### load datasets ########
train_transform_det = trans.Compose([
trans.Scale(cfg.TRANSFROM_SCALES),
])
val_transform_det = trans.Compose([
trans.Scale(cfg.TRANSFROM_SCALES),
])
train_data = data.Dataset(cfg.TRAIN_DATA_PATH,
cfg.TRAIN_LABEL_PATH,
cfg.TRAIN_TXT_PATH,
'train',
transform=True,
transform_med=train_transform_det)
train_loader = Data.DataLoader(train_data,
batch_size=8,
shuffle=True,
num_workers=4,
pin_memory=True)
val_data = data.Dataset(cfg.VAL_DATA_PATH,
cfg.VAL_LABEL_PATH,
cfg.VAL_TXT_PATH,
'val',
transform=True,
transform_med=val_transform_det)
val_loader = Data.DataLoader(val_data,
batch_size=cfg.BATCH_SIZE,
shuffle=False,
num_workers=4,
pin_memory=True)
'''
test_data = dates.Dataset(cfg.TEST_DATA_PATH,cfg.VAL_LABEL_PATH,
cfg.VAL_TXT_PATH,'val',transform=True,
transform_med = val_transform_det)
test_loader = Data.DataLoader(test_data, batch_size= cfg.BATCH_SIZE,
shuffle= False, num_workers= 4, pin_memory= True)
'''
###### build models ########
import model.DASNET as models
G = models.generator
D = models.discriminator
train_model(D, G, train_loader, val_loader)
def get_transform(train):
transforms = []
# converts the image, a PIL image, into a PyTorch Tensor
transforms.append(T.ToTensor())
if train:
# during training, randomly flip the training images
# and ground-truth for data augmentation
transforms.append(T.RandomHorizontalFlip(0.5))
return T.Compose(transforms)
def get_voc(root, image_set, transforms):
t = [ConvertVOCtoCOCO()]
if transforms is not None:
t.append(transforms)
transforms = T.Compose(t)
dataset = VOCDetection(img_folder=root, year='2007', image_set=image_set, transforms=transforms)
return dataset
def train(args):
joint_transform = transforms.Compose([
transforms.RandomScale(),
transforms.Mirror(),
transforms.RandomCrop()
])
trainset = datasets[args.dataset](mode=args.mode, root=args.dataset_root)
net = models[args.g]
def get_transform(train=False):
transforms = []
if train:
transforms.append(custom_T.RandomHorizontalFlip())
transforms.append(custom_T.RandomCrop())
transforms.append(custom_T.ToTensor())
transforms.append(custom_T.FasterRCNNResizer())
return custom_T.Compose(transforms)
def __init__(self,
root: str,
targetHeight: int,
targetWidth: int,
numClass: int,
train: bool = True):
"""
:param root: should contain .jpg files and corresponding .txt files
:param targetHeight: desired height for model input
:param targetWidth: desired width for model input
:param numClass: number of classes in the given dataset
"""
self.cache = {}
imagePaths = glob(os.path.join(root, '*.jpg'))
for path in imagePaths:
name = path.split('/')[-1].split('.jpg')[0]
self.cache[path] = os.path.join(root, f'{name}.txt')
self.paths = list(self.cache.keys())
self.targetHeight = targetHeight
self.targetWidth = targetWidth
self.numClass = numClass
if train:
self.transforms = T.Compose([
T.RandomOrder([
T.RandomHorizontalFlip(),
T.RandomVerticalFlip(),
T.RandomSizeCrop(numClass)
]),
T.Resize((targetHeight, targetWidth)),
T.ColorJitter(brightness=.2, contrast=.1, saturation=.1,
hue=0),
T.Normalize()
])
else:
self.transforms = T.Compose(
[T.Resize((targetHeight, targetWidth)),
T.Normalize()])
def __init__(self, root=None, dataloader=default_loader):
self.transform1 = transforms.Compose([
transforms.RandomRotation(30),
transforms.Resize([256, 256]),
transforms.RandomCrop(INPUT_SIZE),
transforms.RandomHorizontalFlip(),
transforms.ColorJitter(brightness=(0.9, 1.1),
contrast=(0.9, 1.1),
saturation=(0.9, 1.1)),
transforms.ToTensor(),
transforms.Normalize(mean=(0.485, 0.456, 0.406),
std=(0.229, 0.224, 0.225)),
transforms.RandomErasing(probability=0.5, sh=0.05)
])
# 增强方法2： 关注更小的区域
self.transform2 = transforms.Compose([
transforms.RandomRotation(30),
transforms.Resize([336, 336]),
transforms.RandomCrop(INPUT_SIZE),
transforms.RandomHorizontalFlip(),
transforms.ColorJitter(brightness=(0.9, 1.1),
contrast=(0.9, 1.1),
saturation=(0.9, 1.1)),
transforms.ToTensor(),
transforms.Normalize(mean=(0.485, 0.456, 0.406),
std=(0.229, 0.224, 0.225)),
transforms.RandomErasing(probability=0.5, sh=0.05)
])
self.dataloader = dataloader
self.root = root
with open(os.path.join(self.root, TRAIN_DATASET), 'r') as fid:
self.imglist = fid.readlines()
self.labels = []
for line in self.imglist:
image_path, label = line.strip().split()
self.labels.append(int(label))
self.labels = np.array(self.labels)
self.labels = torch.LongTensor(self.labels)
def prepare_transforms(tranform_params):
transforms_list = []
for transform_info in tranform_params:
transform_name = transform_info['name']
transform_params = transform_info['params']
if transform_params is not None:
transform = transforms.__dict__[transform_name](**transform_params)
else:
transform = transforms.__dict__[transform_name]()
transforms_list.append(transform)
transform = transforms.Compose(transforms_list)
return transform
def get_transforms(args):
transforms_list = [
custom_transforms.RandomSizedCrop(args.image_size, args.resize_min,
args.resize_max),
custom_transforms.RandomHorizontalFlip(),
custom_transforms.toTensor()
]
if args.color_space == 'lab':
transforms_list.insert(2, custom_transforms.toLAB())
elif args.color_space == 'rgb':
transforms_list.insert(2, custom_transforms.toRGB('RGB'))
transforms = custom_transforms.Compose(transforms_list)
return transforms
def visualization(x_train, y_train, x_val, y_val, x_test, y_test, nb_class,
ckpt, opt):
# no augmentations used for linear evaluation
transform_lineval = transforms.Compose([transforms.ToTensor()])
train_set_lineval = UCR2018(data=x_train,
targets=y_train,
transform=transform_lineval)
val_set_lineval = UCR2018(data=x_val,
targets=y_val,
transform=transform_lineval)
test_set_lineval = UCR2018(data=x_test,
targets=y_test,
transform=transform_lineval)
train_loader_lineval = torch.utils.data.DataLoader(train_set_lineval,
batch_size=128,
shuffle=True)
val_loader_lineval = torch.utils.data.DataLoader(val_set_lineval,
batch_size=128,
shuffle=False)
test_loader_lineval = torch.utils.data.DataLoader(test_set_lineval,
batch_size=128,
shuffle=False)
signal_length = x_train.shape[1]
# loading the saved backbone
backbone_lineval = SimConv4().cuda() # defining a raw backbone model
checkpoint = torch.load(ckpt, map_location='cpu')
backbone_lineval.load_state_dict(checkpoint)
print('Visualization')
for epoch in range(opt.epochs_test):
backbone_lineval.eval()
embeds = None
labels = None
for i, (data, target) in enumerate(test_loader_lineval):
data = data.cuda()
target = target.cuda().view(-1, 1)
output = backbone_lineval(data).detach()
if embeds is None:
embeds = output.cpu().numpy()
labels = target.cpu().numpy()
else:
embeds = np.vstack([embeds, output.cpu().numpy()])
labels = np.vstack([labels, target.cpu().numpy()])
return embeds, labels
def pretrain_InterSampleRel(x_train, y_train, opt):
K = opt.K
batch_size = opt.batch_size # 128 has been used in the paper
tot_epochs = opt.epochs # 400 has been used in the paper
feature_size = opt.feature_size
ckpt_dir = opt.ckpt_dir
prob = 0.2 # Transform Probability
raw = transforms.Raw()
cutout = transforms.Cutout(sigma=0.1, p=prob)
jitter = transforms.Jitter(sigma=0.2, p=prob)
scaling = transforms.Scaling(sigma=0.4, p=prob)
magnitude_warp = transforms.MagnitudeWrap(sigma=0.3, knot=4, p=prob)
time_warp = transforms.TimeWarp(sigma=0.2, knot=8, p=prob)
window_slice = transforms.WindowSlice(reduce_ratio=0.8, p=prob)
window_warp = transforms.WindowWarp(window_ratio=0.3, scales=(0.5, 2), p=prob)
transforms_list = {'jitter': [jitter],
'cutout': [cutout],
'scaling': [scaling],
'magnitude_warp': [magnitude_warp],
'time_warp': [time_warp],
'window_slice': [window_slice],
'window_warp': [window_warp],
'G0': [jitter, magnitude_warp, window_slice],
'G1': [jitter, time_warp, window_slice],
'G2': [jitter, time_warp, window_slice, window_warp, cutout],
'none': [raw]}
transforms_targets = list()
for name in opt.aug_type:
for item in transforms_list[name]:
transforms_targets.append(item)
train_transform = transforms.Compose(transforms_targets + [transforms.ToTensor()])
backbone = SimConv4().cuda()
model = RelationalReasoning(backbone, feature_size).cuda()
train_set = MultiUCR2018(data=x_train, targets=y_train, K=K, transform=train_transform)
train_loader = torch.utils.data.DataLoader(train_set,
batch_size=batch_size,
shuffle=True)
torch.save(model.backbone.state_dict(), '{}/backbone_init.tar'.format(ckpt_dir))
acc_max, epoch_max = model.train(tot_epochs=tot_epochs, train_loader=train_loader, opt=opt)
torch.save(model.backbone.state_dict(), '{}/backbone_last.tar'.format(ckpt_dir))
return acc_max, epoch_max
def get_dataloader():
# TODO(xwd): Adaptive normalization by some large image.
# E.g. In medical image processing, WSI image is very large and different to ordinary images.
value_scale = 255
mean = [0.485, 0.456, 0.406]
mean = [item * value_scale for item in mean]
std = [0.229, 0.224, 0.225]
std = [item * value_scale for item in std]
train_transform = transform.Compose([
transform.RandomScale([cfg['scale_min'], cfg['scale_max']]),
transform.RandomRotate([cfg['rotate_min'], cfg['rotate_max']],
padding=mean,
ignore_label=cfg['ignore_label']),
transform.RandomGaussianBlur(),
transform.RandomHorizontallyFlip(),
transform.RandomCrop([cfg['train_h'], cfg['train_w']],
crop_type='rand',
padding=mean,
ignore_label=cfg['ignore_label']),
transform.ToTensor(),
transform.Normalize(mean=mean, std=std)
])
train_data = cityscapes.Cityscapes(cfg['data_path'],
split='train',
transform=train_transform)
# Use data sampler to make sure each GPU loads specific parts of dataset to avoid data reduntant.
train_sampler = DistributedSampler(train_data)
train_loader = DataLoader(train_data,
batch_size=cfg['batch_size'] //
cfg['world_size'],
shuffle=(train_sampler is None),
num_workers=4,
pin_memory=True,
sampler=train_sampler,
drop_last=True)
return train_loader, train_sampler
def get_transform(train=False, yolo=False, aug=None):
assert aug == 'dirty_camera_lens' or aug == 'gan' or aug is None, "Aug parameter not valid"
transforms = []
if yolo:
transforms.append(custom_T.PadToSquare())
transforms.append(custom_T.Resize(img_size=None))
if train:
transforms.append(custom_T.RandomHorizontalFlip())
transforms.append(custom_T.RandomCrop())
if aug == 'dirty_camera_lens':
print("Augmentation: Dirty Camera Lens")
transforms.append(custom_T.DirtyCameraLens())
transforms.append(custom_T.ToTensor())
# transforms.append(custom_T.FasterRCNNResizer())
return custom_T.Compose(transforms)
def __init__(self, root=None, dataloader=default_loader):
self.transform = transforms.Compose([
transforms.Resize([256, 256]),
transforms.RandomCrop(INPUT_SIZE),
transforms.ToTensor(),
transforms.Normalize(mean=(0.485, 0.456, 0.406),
std=(0.229, 0.224, 0.225))
])
self.dataloader = dataloader
self.root = root
self.imgs = []
self.labels = []
with open(os.path.join(self.root, EVAL_DATASET), 'r') as fid:
for line in fid.readlines():
img_path, label = line.strip().split()
img = self.dataloader(img_path)
label = int(label)
self.imgs.append(img)
self.labels.append(label)
def deploy(path):
assert os.path.exists(path), f'{path} not found : ('
dataset = 'YOUR_DATASET_NAME'
img_size = 256
test_transform = transforms.Compose([
transforms.Resize((img_size, img_size)),
transforms.ToTensor(),
transforms.Normalize(mean=(0.5, 0.5, 0.5), std=(0.5, 0.5, 0.5))
])
testA = ImageFolder(os.path.join('dataset', dataset, 'testA'), test_transform)
with fluid.dygraph.guard():
testA_loader = DataLoader(testA, batch_size=1, shuffle=False)
real_A, _ = next(iter(testA_loader))
in_np = real_A.numpy()
# load model
place = fluid.CPUPlace()
exe = fluid.Executor(place)
program, feed_vars, fetch_vars = fluid.io.load_inference_model(path, exe)
# inference
fetch, = exe.run(program, feed={feed_vars[0]: in_np}, fetch_list=fetch_vars)
def img_postprocess(img):
assert isinstance(img, np.ndarray), type(img)
img = img * 0.5 + 0.5
img = img.squeeze(0).transpose((1, 2, 0))
# BGR to RGB
img = img[:, :, ::-1]
return img
in_img = img_postprocess(in_np)
out_img = img_postprocess(fetch)
plt.subplot(121)
plt.title('real A')
plt.imshow(in_img)
plt.subplot(122)
plt.title('A to B')
plt.imshow(out_img)
plt.show()
def main():
net = AFENet(classes=21, pretrained_model_path=None).cuda()
net.load_state_dict(torch.load(os.path.join(args['model_save_path'], args['snapshot'])))
value_scale = 255
mean = [0.485, 0.456, 0.406]
mean = [item * value_scale for item in mean]
std = [0.229, 0.224, 0.225]
std = [item * value_scale for item in std]
test_transform = transform.Compose([transform.ToTensor()])
test_data = voc2012.VOC2012(split='test', data_root=args['dataset_root'], data_list=args['test_list'],
transform=test_transform)
test_loader = DataLoader(test_data, batch_size=1, shuffle=False, num_workers=1)
gray_folder = os.path.join(args['test_result_save_path'])
# gray_folder = os.path.join(args['test_result_save_path'], 'gray')
color_folder = os.path.join(args['test_result_save_path'], 'color')
colors = np.loadtxt(args['colors_path']).astype('uint8')
test(test_loader, test_data.data_list, net, 21, mean, std, 512, 480, 480, args['scales'],
gray_folder, color_folder, colors)
def _create_data_loader(img_path, batch_size, img_size, n_cpu):
"""Creates a DataLoader for inferencing.
:param img_path: Path to file containing all paths to validation images.
:type img_path: str
:param batch_size: Size of each image batch
:type batch_size: int
:param img_size: Size of each image dimension for yolo
:type img_size: int
:param n_cpu: Number of cpu threads to use during batch generation
:type n_cpu: int
:return: Returns DataLoader
:rtype: DataLoader
"""
dataset = ImageFolder(img_path,
transform=transforms.Compose(
[DEFAULT_TRANSFORMS,
Resize(img_size)]))
dataloader = DataLoader(dataset,
batch_size=batch_size,
shuffle=False,
num_workers=n_cpu,
pin_memory=True)
return dataloader
#torch.cuda.manual_seed(args.seed)
kwargs = {'num_workers': 1, 'pin_memory': True} if args.cuda else {}
### PARAMETERS ###
# LSTM & Q Learning
IMAGE_SCALE = 28
IMAGE_SIZE = IMAGE_SCALE*IMAGE_SCALE
HIDDEN_LAYERS = 1
HIDDEN_NODES = 200
OUTPUT_CLASSES = args.class_vector_size
##################
train_transform = transforms.Compose([
transforms.Resize((IMAGE_SCALE, IMAGE_SCALE)),
transforms.ToTensor()
])
test_transform = transforms.Compose([
transforms.Resize((IMAGE_SCALE, IMAGE_SCALE)),
transforms.ToTensor()
])
print("Loading trainingsets...")
omniglot_loader = loader.OmniglotLoader('data/omniglot', classify=False, partition=0.8, classes=True)
train_loader = torch.utils.data.DataLoader(
OMNIGLOT('data/omniglot', train=True, transform=train_transform, download=True, omniglot_loader=omniglot_loader, batch_size=args.episode_size),
batch_size=args.mini_batch_size, shuffle=True, **kwargs)
print("Loading testset...")
test_loader = torch.utils.data.DataLoader(
OMNIGLOT('data/omniglot', train=False, transform=test_transform, omniglot_loader=omniglot_loader, batch_size=args.episode_size),
batch_size=args.mini_batch_size, shuffle=True, **kwargs)
def main():
global BEST_LOSS
cudnn.benchmark = True
start_epoch = cfg.OPOSE.start_epoch # start from epoch 0 or last checkpoint epoch
# Create ckpt & vis folder
if not os.path.isdir(cfg.OPOSE.ckpt):
os.makedirs(cfg.OPOSE.ckpt)
if not os.path.exists(os.path.join(cfg.OPOSE.ckpt, 'vis')):
os.makedirs(os.path.join(cfg.OPOSE.ckpt, 'vis'))
if args.cfg_file is not None and not cfg.OPOSE.evaluate:
shutil.copyfile(
args.cfg_file,
os.path.join(cfg.OPOSE.ckpt,
args.cfg_file.split('/')[-1]))
model = pose_estimation.PoseModel(num_point=19,
num_vector=19,
pretrained=True)
# # Calculate FLOPs & Param
# n_flops, n_convops, n_params = measure_model(model, cfg.OPOSE.input_size, cfg.OPOSE.input_size)
criterion = nn.MSELoss().cuda()
# Dataset and Loader
train_dataset = dataset.CocoOpenposeData(
cfg,
cfg.OPOSE.data_root,
cfg.OPOSE.info_root,
'train2017',
transformer=transforms.Compose([
transforms.RandomResized(),
transforms.RandomRotate(40),
transforms.RandomCrop(368),
transforms.RandomHorizontalFlip(),
]))
train_loader = torch.utils.data.DataLoader(train_dataset,
batch_size=BATCH_SIZE,
shuffle=True,
num_workers=int(cfg.workers),
pin_memory=True)
if cfg.OPOSE.validate or cfg.OPOSE.evaluate:
val_dataset = dataset.CocoOpenposeData(
cfg,
cfg.OPOSE.data_root,
cfg.OPOSE.info_root,
'val2017',
transformer=transforms.Compose(
[transforms.TestResized(cfg.OPOSE.input_size)]))
val_loader = torch.utils.data.DataLoader(val_dataset,
batch_size=BATCH_SIZE,
shuffle=False,
num_workers=int(cfg.workers),
pin_memory=True)
# Load nets into gpu
if NUM_GPUS > 1:
model = torch.nn.DataParallel(model, device_ids=gpu).cuda()
# Set up optimizers
params, multiple = get_parameters(model, cfg, False)
optimizer = torch.optim.SGD(params,
cfg.OPOSE.base_lr,
momentum=cfg.OPOSE.momentum,
weight_decay=cfg.OPOSE.weight_decay)
# Resume training
title = 'Pytorch-OPOSE-{}-{}'.format(cfg.OPOSE.arch_encoder,
cfg.OPOSE.arch_decoder)
if cfg.OPOSE.resume:
# Load checkpoint.
print("==> Resuming from checkpoint '{}'".format(cfg.OPOSE.resume))
assert os.path.isfile(
cfg.OPOSE.resume), 'Error: no checkpoint directory found!'
ckpt = torch.load(cfg.OPOSE.resume)
BEST_LOSS = ckpt['best_loss']
start_epoch = ckpt['epoch']
try:
model.module.load_state_dict(ckpt['state_dict'])
except:
model.load_state_dict(ckpt['state_dict'])
optimizer.load_state_dict(ckpt['optimizer'])
logger = Logger(os.path.join(cfg.OPOSE.ckpt, 'log.txt'),
title=title,
resume=True)
else:
logger = Logger(os.path.join(cfg.OPOSE.ckpt, 'log.txt'), title=title)
logger.set_names(
['epoch', 'Learning Rate', 'Train Loss', 'Valid Loss'])
# Train and val
for epoch in range(start_epoch, cfg.OPOSE.epochs):
print('\nEpoch: [{}/{}] | LR: {:.8f} '.format(epoch + 1,
cfg.OPOSE.epochs,
cfg.OPOSE.base_lr))
train_loss = train(train_loader, model, criterion, optimizer, epoch,
USE_CUDA)
if cfg.OPOSE.validate:
test_loss = test(val_loader, model, criterion, optimizer, epoch,
USE_CUDA)
else:
test_loss = 0.0, 0.0
# Append logger file
logger.append([epoch, cfg.OPOSE.base_lr, train_loss, test_loss])
# Save model
save_checkpoint(model, optimizer, test_loss, epoch)
# Adjust learning rate
adjust_learning_rate(optimizer, epoch)
# Draw curve
try:
draw_curve('model', cfg.OPOSE.ckpt)
print('==> Success saving log curve...')
except:
print('==> Saving log curve error...')
logger.close()
try:
savefig(os.path.join(cfg.OPOSE.ckpt, 'log.eps'))
shutil.copyfile(
os.path.join(cfg.OPOSE.ckpt, 'log.txt'),
os.path.join(
cfg.OPOSE.ckpt, 'log{}.txt'.format(
datetime.datetime.now().strftime('%Y%m%d%H%M%S'))))
except:
print('Copy log error.')
print('==> Training Done!')
print('==> Best acc: {:.4f}%'.format(BEST_LOSS))