def train(net):
net.train()
priorbox = PriorBox()
with torch.no_grad():
priors = priorbox.forward()
priors = priors.to(device)
dataloader = DataLoader(VOCDetection(), batch_size=2, collate_fn=detection_collate, num_workers=12)
for epoch in range(1000):
loss_ls, loss_cs = [], []
load_t0 = time.time()
if epoch > 500:
adjust_learning_rate(optimizer, 1e-4)
for images, targets in dataloader:
images = images.to(device)
targets = [anno.to(device) for anno in targets]
out = net(images)
optimizer.zero_grad()
loss_l, loss_c = criterion(out, priors, targets)
loss = 2 * loss_l + loss_c
loss.backward()
optimizer.step()
loss_cs.append(loss_c.item())
loss_ls.append(loss_l.item())
load_t1 = time.time()
print(f'{np.mean(loss_cs)}, {np.mean(loss_ls)} time:{load_t1-load_t0}')
torch.save(net.state_dict(), 'Final_FaceBoxes.pth')
def train(net):
net.train()
priorbox = PriorBox()
with torch.no_grad():
priors = priorbox.forward()
priors = priors.to(device)
dataloader = DataLoader(VOCDetection(),
batch_size=2,
collate_fn=detection_collate,
num_workers=12)
for epoch in range(1000):
loss_ls, loss_cs = [], []
load_t0 = time.time()
if epoch > 500:
adjust_learning_rate(optimizer, 1e-4)
for images, targets in dataloader:
images = images.to(device)
targets = [anno.to(device) for anno in targets]
out = net(images)
optimizer.zero_grad()
loss_l, loss_c = criterion(out, priors, targets)
loss = 2 * loss_l + loss_c
loss.backward()
optimizer.step()
loss_cs.append(loss_c.item())
loss_ls.append(loss_l.item())
load_t1 = time.time()
print(f'{np.mean(loss_cs)}, {np.mean(loss_ls)} time:{load_t1-load_t0}')
torch.save(net.state_dict(), 'Final_FaceBoxes.pth')
def init_priors(self, feature_maps, image_size):
# Hacky key system, but works....
key = ".".join([str(item) for i in range(len(feature_maps)) for item in feature_maps[i]]) + \
"," + ".".join([str(_) for _ in image_size])
if key in self.prior_cache:
return self.prior_cache[key].clone()
priorbox = PriorBox(self.cfg, image_size, feature_maps)
prior = priorbox.forward()
self.prior_cache[key] = prior.clone()
return prior
def __init__(self, phase, size, Backbone, Neck, Head, cfg):
super(SSD, self).__init__()
self.phase = phase
self.cfg = cfg
self.priorbox = PriorBox(self.cfg)
self.priors = self.priorbox.forward()
self.size = size
# SSD network
self.backbone = Backbone
self.neck = Neck
self.head = Head
self.num_classes = cfg['num_classes']
self.softmax = nn.Softmax(dim=-1)
self.detect = Detect(self.num_classes , 0, 200, 0.01, 0.45,variance = cfg['variance'], nms_kind=cfg['nms_kind'], beta1=cfg['beta1'])
def __init__(self, model_path, gpu_ids, layers, score_thresh=0.5):
"""
检测整体基本流程
:param model_path: 模型路径
:param gpu_ids: gpu序列号
:param layers: 18 , 50
:param score_thresh: 置信度过滤
"""
self.keep_top_k = 100
self.nms_threshold = 0.3
self.nms_score = score_thresh
self.nms_threshold = self.nms_threshold
self.test_size = 640
self.__model_path = model_path
self.__gpu_ids = gpu_ids
self.device = torch.device('cuda:{}'.format(str(gpu_ids)) if torch.
cuda.is_available() else 'cpu')
self.layers = layers
self.model = RetinaFace(self.layers)
self.model = self.__load_model(self.model, self.__model_path)
self.model = self.model.to(self.device)
self.model.eval()
self.priorbox = PriorBox(box_specs_list=[[(0.6, 0.5), (0.75, 1.),
(0.9, 1.)],
[(0.2, 0.5), (0.4, 1.),
(0.6, 1.)],
[(0.05, 0.5), (0.1, 1.),
(0.2, 1.)],
[(0.0125, 0.5), (0.025, 1.),
(0.05, 1.)]],
base_anchor_size=[1.0, 1.0])
self.priors = self.priorbox.generate(feature_map_shape_list=[(10, 10),
(20, 20),
(40, 40),
(80, 80)],
im_height=640,
im_width=640)
self.priors = self.priors.to(self.device)
self.mean = torch.Tensor([104, 117, 123]).to(self.device)
self.variance = torch.Tensor([0.1, 0.2]).to(self.device)
self.Decode = Decode(self.priors.data, self.variance)
def __init__(self, confidence_threshold=0.02, top_k=1000, nms_threshold=0.4, keep_top_k=500, vis_thres=0.6):
self.net = Retina(cfg=cfg_mnet).to(device)
self.net = load_model(self.net, 'mnet_plate.pth', False)
self.net.eval()
self.lprnet = plate_recogition()
self.priorbox = PriorBox(cfg_mnet)
self.confidence_threshold = confidence_threshold
self.top_k = top_k
self.nms_threshold = nms_threshold
self.keep_top_k = keep_top_k
self.vis_thres = vis_thres
self.resize = 1
self.points_ref = np.float32([[0, 0], [94, 0], [0, 24], [94, 24]])
class SSD(nn.Module):
"""Single Shot Multibox Architecture
The network is composed of a base VGG network followed by the
added multibox conv layers. Each multibox layer branches into
1) conv2d for class conf scores
2) conv2d for localization predictions
3) associated priorbox layer to produce default bounding
boxes specific to the layer's feature map size.
See: https://arxiv.org/pdf/1512.02325.pdf for more details.
Args:
phase: (string) Can be "test" or "train"
size: input image size
base: VGG16 layers for input, size of either 300 or 500
extras: extra layers that feed to multibox loc and conf layers
head: "multibox head" consists of loc and conf conv layers
"""
def __init__(self, phase, size, Basenet, Neck, Head, cfg):
super(SSD, self).__init__()
self.phase = phase
self.cfg = cfg
self.priorbox = PriorBox(self.cfg)
self.priors = self.priorbox.forward()
self.size = size
# SSD network
self.basenet = Basenet
self.neck = Neck
self.head = Head
self.num_classes = cfg['num_classes']
self.softmax = nn.Softmax(dim=-1)
self.detect = Detect(self.num_classes, 0, 200, 0.01, 0.45,
variance=cfg['variance'], nms_kind=cfg['nms_kind'], beta1=cfg['beta1'])
def forward(self, x, phase):
"""Applies network layers and ops on input image(s) x.
Args:
x: input image or batch of images. Shape: [batch,3,300,300].
Return:
Depending on phase:
test:
Variable(tensor) of output class label predictions,
confidence score, and corresponding location predictions for
each object detected. Shape: [batch,topk,7]
train:
list of concat outputs from:
1: confidence layers, Shape: [batch*num_priors,num_classes]
2: localization layers, Shape: [batch,num_priors*4]
3: priorbox layers, Shape: [2,num_priors*4]
"""
x = self.basenet(x)
if self.neck is not None:
x = self.neck(x)
conf, loc = self.head(x)
loc = torch.cat([o.view(o.size(0), -1) for o in loc], 1)
conf = torch.cat([o.view(o.size(0), -1) for o in conf], 1)
if phase == "test":
output = self.detect.trans(
loc.view(loc.size(0), -1, 4), # loc preds
self.softmax(conf.view(conf.size(0), -1,
self.num_classes)), # conf preds
#self.priors.type(type(x.data)) # default boxes
self.priors
)
else:
output = (
loc.view(loc.size(0), -1, 4),
conf.view(conf.size(0), -1, self.num_classes),
self.priors
)
return output
def load_weights(self, base_file):
other, ext = os.path.splitext(base_file)
if ext == '.pkl' or '.pth':
print('Loading weights into state dict...')
self.load_state_dict(torch.load(base_file,
map_location=lambda storage, loc: storage))
print('Finished!')
else:
print('Sorry only .pth and .pkl files supported.')
print('Loading Dataset...')
(show_classes, num_classes, dataset, epoch_size, max_iter, testset) = load_dataset()
print('Loading Network...')
from models.detector import Detector
model = Detector(args.size, num_classes, args.backbone, args.neck)
model.train()
model.cuda()
num_param = sum(p.numel() for p in model.parameters() if p.requires_grad)
print('Total param is : {:e}'.format(num_param))
print('Preparing Optimizer & AnchorBoxes...')
optimizer = optim.SGD(tencent_trick(model), lr=args.lr, momentum=0.9, weight_decay=0.0005)
criterion = MultiBoxLoss(num_classes, mutual_guide=args.mutual_guide)
priorbox = PriorBox(args.base_anchor_size, args.size)
with torch.no_grad():
priors = priorbox.forward()
priors = priors.cuda()
if args.trained_model is not None:
print('loading weights from', args.trained_model)
state_dict = torch.load(args.trained_model)
model.load_state_dict(state_dict, strict=True)
else:
print('Training {}-{} on {} with {} images'.format(args.neck, args.backbone, dataset.name, len(dataset)))
os.makedirs(args.save_folder, exist_ok=True)
epoch = 0
timer = Timer()
for iteration in range(max_iter):
if iteration % epoch_size == 0:
class FaceDetector(object):
def __init__(self, model_path, gpu_ids, layers, score_thresh=0.5):
"""
检测整体基本流程
:param model_path: 模型路径
:param gpu_ids: gpu序列号
:param layers: 18 , 50
:param score_thresh: 置信度过滤
"""
self.keep_top_k = 100
self.nms_threshold = 0.3
self.nms_score = score_thresh
self.nms_threshold = self.nms_threshold
self.test_size = 640
self.__model_path = model_path
self.__gpu_ids = gpu_ids
self.device = torch.device('cuda:{}'.format(str(gpu_ids)) if torch.
cuda.is_available() else 'cpu')
self.layers = layers
self.model = RetinaFace(self.layers)
self.model = self.__load_model(self.model, self.__model_path)
self.model = self.model.to(self.device)
self.model.eval()
self.priorbox = PriorBox(box_specs_list=[[(0.6, 0.5), (0.75, 1.),
(0.9, 1.)],
[(0.2, 0.5), (0.4, 1.),
(0.6, 1.)],
[(0.05, 0.5), (0.1, 1.),
(0.2, 1.)],
[(0.0125, 0.5), (0.025, 1.),
(0.05, 1.)]],
base_anchor_size=[1.0, 1.0])
self.priors = self.priorbox.generate(feature_map_shape_list=[(10, 10),
(20, 20),
(40, 40),
(80, 80)],
im_height=640,
im_width=640)
self.priors = self.priors.to(self.device)
self.mean = torch.Tensor([104, 117, 123]).to(self.device)
self.variance = torch.Tensor([0.1, 0.2]).to(self.device)
self.Decode = Decode(self.priors.data, self.variance)
def detect(self, detect_input):
images, percent = self.preprocess(detect_input) #前处理
loc, conf, landms = self.inference(images) #推理
boxes, landms, scores = self.decode(loc, conf, landms, percent) #解码
boxes, landms, scores = self.postprocess(boxes, landms, scores) #后处理
if len(scores) == 0:
return None, None, None
else:
return boxes, landms, scores
def __check_keys(self, model, pretrained_state_dict):
ckpt_keys = set(pretrained_state_dict.keys())
model_keys = set(model.state_dict().keys())
used_pretrained_keys = model_keys & ckpt_keys
unused_pretrained_keys = ckpt_keys - model_keys
missing_keys = model_keys - ckpt_keys
print('Missing keys:{}'.format(len(missing_keys)))
print('Unused checkpoint keys:{}'.format(len(unused_pretrained_keys)))
print('Used keys:{}'.format(len(used_pretrained_keys)))
assert len(
used_pretrained_keys) > 0, 'load NONE from pretrained checkpoint'
return True
def __remove_prefix(self, state_dict, prefix):
''' Old style model is stored with all names of parameters sharing common prefix 'module.' '''
print('remove prefix \'{}\''.format(prefix))
f = lambda x: x.split(prefix, 1)[-1] if x.startswith(prefix) else x
return {f(key): value for key, value in state_dict.items()}
def __load_model(self, model, model_path):
print('Loading pretrained model from {}'.format(model_path))
if self.__gpu_ids == None:
device = torch.cuda.current_device()
pretrained_dict = torch.load(
model_path,
map_location=lambda storage, loc: storage.cuda(device))
else:
pretrained_dict = torch.load(
model_path, map_location=lambda storage, loc: storage)
if "state_dict" in pretrained_dict.keys():
pretrained_dict = self.__remove_prefix(
pretrained_dict['state_dict'], 'module.')
else:
pretrained_dict = self.__remove_prefix(pretrained_dict, 'module.')
self.__check_keys(model, pretrained_dict)
model.load_state_dict(pretrained_dict, strict=False)
return model
def preprocess(self, image):
"""Preprocess"""
image_t, percent = no_deform_resize_pad(image, self.test_size)
image_t = rgb_mean_gpu(image_t, self.mean, self.device)
images = image_t.permute(2, 0, 1).reshape(1, 3, self.test_size,
self.test_size)
return images, percent
def inference(self, images):
"""网络推理"""
with torch.no_grad():
loc, conf, landms = self.model(images)
return loc, conf, landms
def decode(self, loc, conf, landms, percent):
"""推理结果的解码"""
boxes = self.Decode.decode_bbox(loc.squeeze(0).data)
landms = self.Decode.decode_landm(landms.squeeze(0).data)
detect_boxes = boxes * self.test_size * percent
detect_landmas = landms * self.test_size * percent
scores = conf.squeeze(0).data[:, 1]
return detect_boxes, detect_landmas, scores
def postprocess(self, boxes, landms, scores):
"""后处理NMS"""
inds = torch.where(scores >= self.nms_score)[0]
boxes = boxes[inds]
scores = scores[inds]
landms = landms[inds]
keep = nms(boxes, scores, self.nms_threshold)
boxes = boxes[keep]
scores = scores[keep]
landms = landms[keep]
boxes = boxes[:self.keep_top_k, :].cpu().numpy()
scores = scores[:self.keep_top_k].cpu().numpy()
landms = landms[:self.keep_top_k, :].cpu().numpy()
return boxes, landms, scores
"output3": {0: "batch_size"}}
torch.onnx.export(net, dummy_input, onnx_output, verbose=True,
input_names=input_names,
output_names=output_names,
opset_version=12,
dynamic_axes=dynamic_axes)
if False:
model = onnx.load(onnx_output)
model_simp, check = simplify(model)
assert check, "Simplified ONNX model could not be validated"
output_path = 'simp.onnx'
onnx.save(model_simp, output_path)
print('finished exporting onnx ')
img_path = 'export/028125-87_110-204&496_524&585-506&564_204&585_210&514_524&496-0_0_5_24_29_33_24_24-52-45.jpg'
priorbox = PriorBox(cfg_mnet)
points_ref = np.float32([[0, 0], [94, 0], [0, 24], [94, 24]])
confidence_threshold=0.02
top_k=1000
nms_threshold=0.4
keep_top_k=500
vis_thres=0.6
srcimg = cv2.imread(img_path)
img = srcimg.astype('float32')
im_height, im_width, _ = img.shape
img -= (104, 117, 123)
with torch.no_grad():
scale = torch.Tensor([img.shape[1], img.shape[0], img.shape[1], img.shape[0]]).to(device)
img = torch.from_numpy(img).permute(2, 0, 1).unsqueeze(0).to(device)
loc, conf, landms = net(img) # forward pass
