Created by: Chgtaxihe

使用dev分支的opt工具转换模型时，出现以下提示:

[F  2/26 23:39:57.188 ...ite/core/mir/type_precision_cast_pass.cc:133 AddCastInst] Check failed: is_found: Can't find a Cast kernel for Cast op: Tensor<arm,int64_t,NCHW,0>:fill_constant_0.tmp_0->Tensor<arm,float,NCHW,0>:less_than

进一步尝试后发现加操作亦不能转换

Can't find a Cast kernel for Cast op: Tensor<arm,int64_t,NCHW,0>:fill_constant_0.tmp_0->Tensor<arm,float,NCHW,0>:elementwise_add

希望Paddle-Lite能够添加对int64运算的支持