pretrained networks for deep learning applications

Pretrained networks for a variety of problems. WBIR networks are named wbir* put them all in a folder called "models" then see create_wbir_models.R for how to load them wbir_random_vgg_3d --- 3dvgg initialized with random weights wbir_resnet_vgg_weights --- resnet trained to mimic 2d pseudo vgg activations wbir_vggtrue3d -- 3d egg trained to mimic 2d pseudo vgg activations wbir_brain_age_weights -- weights for brain age wbir_pseudo_vgg_ -- 2d egg transferred to 3d

适用于多种任务的预训练神经网络。WBIR网络的命名格式为wbir*，需将所有此类网络存放至名为"models"的文件夹中，随后可参考create_wbir_models.R文件加载这些网络。
wbir_random_vgg_3d --- 采用随机权重初始化的3D VGG神经网络
wbir_resnet_vgg_weights --- 经训练以拟合二维伪VGG激活值的ResNet网络
wbir_vggtrue3d --- 经训练以拟合二维伪VGG激活值的3D EGG模型
wbir_brain_age_weights --- 脑年龄预测任务专用权重参数
wbir_pseudo_vgg_ --- 迁移至三维架构的二维伪VGG EGG模型