Add detail to type for functional.to_tensor pic #8449
Conversation
This is to avoid the pyright error `reportUnknownMemberType` when using this function.
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/vision/8449
Note: Links to docs will display an error until the docs builds have been completed. This comment was automatically generated by Dr. CI and updates every 15 minutes. |
torchvision/transforms/functional.py
Outdated
| @@ -124,7 +124,7 @@ def _is_numpy_image(img: Any) -> bool: | |||
| return img.ndim in {2, 3} | |||
|
|
|||
|
|
|||
| def to_tensor(pic: Union[PILImage, np.ndarray]) -> Tensor: | |||
| def to_tensor(pic: Union[PILImage, np.ndarray[int, int]]) -> Tensor: | |||
There was a problem hiding this comment.
@adamtheturtle thanks for the PR - just to make sure I understand correctly, does np.ndarray[int, int] mean the ndarry is expected to only have 2 dimensions? If that's the case then that is not correct as we allow both 2D and 3D arrays.
There was a problem hiding this comment.
Apologies @NicolasHug for leaving this hanging. I have had another go, and not specified the dimensions. I have also used numpy.typing.NDArray.
|
Thank you for the initial review @NicolasHug . I'd appreciate another look, and I can be quicker with responses from now. |
torchvision/transforms/functional.py
Outdated
| @@ -124,7 +125,7 @@ def _is_numpy_image(img: Any) -> bool: | |||
| return img.ndim in {2, 3} | |||
|
|
|||
|
|
|||
| def to_tensor(pic: Union[PILImage, np.ndarray]) -> Tensor: | |||
| def to_tensor(pic: Union[PILImage, NDArray[np.uint8]]) -> Tensor: | |||
There was a problem hiding this comment.
Sorry @adamtheturtle I don't think this is correct, to_tensor also accepts numpy arrays of float values.
There was a problem hiding this comment.
Thank you for taking another look @NicolasHug . I have attempted to update this PR. We will get there!
This is to avoid the pyright error
reportUnknownMemberTypewhen using this function.