### 算子说明 switch(tensor<int64|int32|int8|bool> cases,listtensor<any> Xs)->(tensor<any> Y) cases、 X[i]、 Y 这三者的shape必须相同 where算子,复用switch where(tensor<int64|int32|int8|bool> cases,tensor<any> A,tensor<any> B)->(tensor<any> Y) ### excuter - cuda - ompsimd ### python