基于飞桨框架的UpsamplingNearest2D的复现及验证

Retraction and verification of UpsamplingNearest2D based on flying paddle frame

5G技术的发展,计算机视觉的应用场景得到了空前的发展。随着飞桨框架的不断完善,越来越多的开发者开始使用飞桨框架。为了保证百度飞桨框架的正确性与稳定性,文章复现了paddle框架paddle.nn.UpsamplingNearest2D并从参数覆盖、正确性验证、数据类型覆盖、异常输入等方面对UpsamplingNearest2D进行了单元测试。

The development of 5G technology, the application scene of computer vision has been an unprecedented development.As the flying paddle frame continues to improve, more and more developers began to use the flying paddle frame. In order to ensure the correctness and stability of Baidu paddle framework, the article reproduces the paddle framework paddle.nn. UpsamplingNearest2D and tested UpsamplingNearest2D from the aspects of parameter coverage, correctness verification, data type coverage and abnormal input.