A bug in module for GPU

[rose-jinyang]

Hello
How are you?
Thanks for contributing to this project.
I found a bug in module for GPU.

The script that I ran is the following:

import paddle
import paddle.nn.functional as F

a = paddle.randint(0, 10, (4, 4))
b = F.one_hot(a, 5)
print(a)
print(b)

The above script does NOT yield any issue when running on GPU.
But when running on CPU it yields an issue.

[paddle-bot-old]

您好，我们已经收到了您的问题，会安排技术人员尽快解答您的问题，请耐心等待。请您再次检查是否提供了清晰的问题描述、复现代码、环境&版本、报错信息等。同时，您也可以通过查看官网API文档、常见问题、历史Issue、AI社区来寻求解答。祝您生活愉快～

Hi! We've received your issue and please be patient to get responded. We will arrange technicians to answer your questions as soon as possible. Please make sure that you have posted enough message to demo your request. You may also check out the API，FAQ，Github Issue and AI community to get the answer.Have a nice day!

[zhiboniu]

The input Tensor's element(0<x<10 here) should not bigger than class_num in API one_hot（which is 5 here）。
For more details please refer to:
https://www.paddlepaddle.org.cn/documentation/docs/en/api/paddle/nn/functional/one_hot_en.html

[rose-jinyang]

That's right.
However, the problem is that the above script does NOT yield any issue on GPU.

[zhiboniu]

O, that is not right action on GPU, thanks for your information, we will fix this bug soonly.