Created by: Archimondecy

传入数据：inputs['_y'] = fluid.layers.data(name="_y", shape=[1], dtype="float32", lod_level=0) 训练计算loss： cost = fluid.layers.square_error_cost(input=pos_score, label=inputs["_y"]) 此处报错，说_y 是double，不是float，求帮忙解决 Tensor holds the wrong type, it holds double, but desires to be float