def generate(self):
model_path = os.path.expanduser(self.model_path)
assert model_path.endswith('.h5'), 'Keras model or weights must be a .h5 file.'
self.model = facenet(self.input_shape, backbone=self.backbone, mode="predict")
self.model.load_weights(self.model_path, by_name=True)
self.model.summary()
print('{} model, anchors, and classes loaded.'.format(model_path))
def generate(self):
#---------------------------------------------------#
# 载入模型与权值
#---------------------------------------------------#
model_path = os.path.expanduser(self.model_path)
assert model_path.endswith(
'.h5'), 'Keras model or weights must be a .h5 file.'
self.model = facenet(self.input_shape,
backbone=self.backbone,
mode="predict")
print('Loading weights into state dict...')
self.model.load_weights(self.model_path, by_name=True)
print('{} model loaded.'.format(model_path))
def generate(self):
# 载入模型
print('Loading weights into state dict...')
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
model = facenet(backbone=self.backbone, mode="predict")
model.load_state_dict(torch.load(self.model_path, map_location=device),
strict=False)
self.net = model.eval()
if self.cuda:
self.net = torch.nn.DataParallel(self.net)
cudnn.benchmark = True
self.net = self.net.cuda()
print('{} model loaded.'.format(self.model_path))
lfw_dir_path = "lfw"
lfw_pairs_path = "model_data/lfw_pair.txt"
#------------------------------------------------------#
# 设置用到的显卡
#------------------------------------------------------#
os.environ["CUDA_VISIBLE_DEVICES"] = ','.join(str(x) for x in train_gpu)
ngpus_per_node = len(train_gpu)
print('Number of devices: {}'.format(ngpus_per_node))
num_classes = get_num_classes(annotation_path)
#---------------------------------#
# 载入模型并加载预训练权重
#---------------------------------#
model_body = facenet(input_shape,
num_classes,
backbone=backbone,
mode="train")
if model_path != '':
#------------------------------------------------------#
# 载入预训练权重
#------------------------------------------------------#
print('Load weights {}.'.format(model_path))
model_body.load_weights(model_path, by_name=True, skip_mismatch=True)
if ngpus_per_node > 1:
model = multi_gpu_model(model_body, gpus=ngpus_per_node)
else:
model = model_body
#-------------------------------------------------------#
# 0.01用于验证，0.99用于训练
#-------------------------------------------------------#
#--------------------------------------#
input_shape = [160, 160, 3]
#--------------------------------------#
# 训练好的权值文件
#--------------------------------------#
model_path = "model_data/facenet_mobilenet.h5"
#--------------------------------------#
# LFW评估数据集的文件路径
# 以及对应的txt文件
#--------------------------------------#
lfw_dir_path = "lfw"
lfw_pairs_path = "model_data/lfw_pair.txt"
#--------------------------------------#
# 评估的批次大小和记录间隔
#--------------------------------------#
batch_size = 256
log_interval = 1
#--------------------------------------#
# ROC图的保存路径
#--------------------------------------#
png_save_path = "model_data/roc_test.png"
test_loader = LFWDataset(dir=lfw_dir_path,
pairs_path=lfw_pairs_path,
batch_size=batch_size,
input_shape=input_shape)
model = facenet(input_shape, backbone=backbone, mode="predict")
model.load_weights(model_path, by_name=True)
test(test_loader, model, png_save_path, log_interval, batch_size)
#--------------------------------------------#
# 该部分代码只用于看网络结构，并非测试代码
#--------------------------------------------#
import os
from nets.facenet import facenet
if __name__ == "__main__":
input_shape = [160, 160, 3]
model = facenet(input_shape,
len(os.listdir("./datasets")),
backbone="inception_resnetv1",
mode="train")
model.summary()
for i, layer in enumerate(model.layers):
print(i, layer.name)
#--------------------------------------------#
# 该部分代码只用于看网络结构，并非测试代码
#--------------------------------------------#
from nets.facenet import facenet
from utils.utils import net_flops
if __name__ == "__main__":
input_shape = [160, 160]
backbone = "mobilenet"
model = facenet([input_shape[0], input_shape[1], 3],
10575,
backbone=backbone,
mode="train")
#--------------------------------------------#
# 查看网络结构网络结构
#--------------------------------------------#
model.summary()
#------------------------------------------#
# 计算网络的FLOPS
#--------------------------------------------#
net_flops(model, table=False)
#--------------------------------------------#
# 获得网络每个层的名称与序号
#--------------------------------------------#
# for i,layer in enumerate(model.layers):
# print(i,layer.name)