由于形状,我确实有将图像输入数据发送到张量流模型的问题,我不知道如何在golang中设置“无”批处理大小
tensorflow-golang | 2018/09/20 19:20:59 main.go:139:形状:[281 50 50 1],类型:1 张量流-golang | 2018/09/20 19:20:59 main.go:157:您必须使用dtype float和shape [?,50,50,1]输入占位符张量'Placeholder'的值 张量流-golang | [[节点:占位符= Placeholder_output_shapes = [[?, 50,50,1]],dtype = DT_FLOAT,shape = [?, 50,50,1],_ device =“ / job:localhost / replica:0 / task: 0 /设备:CPU:0“]]