ViG核心代码及网络结构图

ti_vig

def pvig_ti_224_gelu(pretrained=False, **kwargs):
    class OptInit:
        def __init__(self, num_classes=1000, drop_path_rate=0.0, **kwargs):
            self.k = 9 # 邻域的数目，默认为9
            self.conv = 'mr' # 图卷积层=mr
            self.act = 'gelu' # 激活层=gelu
            self.norm = 'batch' # batch or instance normalization {batch, instance}
            self.bias = True # bias of conv layer True or False
            self.dropout = 0.0 # dropout rate
            self.use_dilation = True # use dilated knn or not
            self.epsilon = 0.2 # stochastic epsilon for gcn
            self.use_stochastic = False # stochastic for gcn, True or False
            self.drop_path = drop_path_rate
            self.blocks = [2,2,6,2] # number of basic blocks in the backbone
            self.channels = [48, 96, 240, 384] # number of channels of deep features
            self.n_classes = num_classes # Dimension of out_channels
            self.emb_dims = 1024 # Dimension of embeddings

    opt = OptInit(**kwargs)
    model = DeepGCN(opt)
    model.default_cfg = default_cfgs['vig_224_gelu']
    return model

DeepGCN

class DeepGCN(torch.nn.Module):
    def __init__(self, opt):
        super(DeepGCN, self).__init__()
        print(opt)
        k = opt