Skip to content

[API Compatibility] Add alias paddle.random.get_rng_state #78472

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
zhwesky2010 merged 4 commits into from
Apr 10, 2026
Merged

[API Compatibility] Add alias paddle.random.get_rng_state #78472

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
6dbc442
add alias paddle.random.get_rng_state
algorithm1832 Mar 25, 2026
e6a59ad
move implementation to paddle.random
algorithm1832 Apr 4, 2026
a9be3b9
fix test
algorithm1832 Apr 8, 2026
92d2dcd
update alias target to paddle.device.cpu.get_rng_state
algorithm1832 Apr 9, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion python/paddle/framework/random.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -71,7 +71,7 @@ def seed(seed: int) -> paddle.base.core.Generator:

def get_rng_state(
Copy link
Copy Markdown
Contributor

@zhwesky2010 zhwesky2010 Apr 8, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

看下torch源码，torch.random.get_rng_state 没有device参数输入，torch.random.*更偏向于cpu的实现，而torch.cuda.random.*是gpu的实现。

这个更适合从Paddle/python/paddle/device/cpu.py导入，并移除无用的device参数，注意最终paddle.random.get_rng_state与torch.random.get_rng_state需要对齐。

Copy link
Copy Markdown
Contributor Author

@algorithm1832 algorithm1832 Apr 9, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

device参数不太能在原地移除吧，在python\paddle\device\__init__.py中，通过判断平台支持的设备类型自动选择get_rng_state，如果移除device参数会导致方法签名不一样

torch没有这个device参数，直接转换前缀，则device为默认值None，应该不会有问题

device: str | None = None,
) -> list[paddle.base.core.GeneratorState]:
) -> list[core.GeneratorState]:
"""
Get all random states of random generators of specified device.

Expand Down
1 change: 1 addition & 0 deletions python/paddle/random.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -15,6 +15,7 @@
from __future__ import annotations

from paddle.base import core
from paddle.device.cpu import get_rng_state # noqa: F401

__all__ = ["initial_seed"]

Expand Down
5 changes: 5 additions & 0 deletions test/compat/test_rng_state.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -71,6 +71,11 @@ def test_invalid_device_raises(self):
finally:
paddle.set_rng_state(original_state)

def test_api_compatibility(self):
paddle_obj = paddle.device.cpu.get_rng_state
alias_obj = paddle.random.get_rng_state
self.assertTrue(paddle_obj is alias_obj)


if __name__ == "__main__":
unittest.main()
Loading