Add fp16 mul op support and bind paddle fp16 to numpy fp16 (#9017)
* add fp16 mul op support * small fix * fix bug * small fix * fix PADDLE_WITH_CUDA compiling issue * reorg code * test for pybind * treate as float16 as uint16_t in pybind * bind np.float16 to paddle float16 * small fix * clean code * remove redundancy * fix mul_op test * address comments * small fix * add is_float16_supported func
Showing
想要评论请 注册 或 登录