[proto] Ported LinearTransformation #6458
Conversation
| shape = inpt.shape | ||
| n = shape[-3] * shape[-2] * shape[-1] | ||
| if n != self.transformation_matrix.shape[0]: | ||
| raise ValueError( | ||
| "Input tensor and transformation matrix have incompatible shape." | ||
| + f"[{shape[-3]} x {shape[-2]} x {shape[-1]}] != " | ||
| + f"{self.transformation_matrix.shape[0]}" | ||
| ) |
There was a problem hiding this comment.
Shouldn't this happen in _get_params?
There was a problem hiding this comment.
There are no params to generate in this transform. I think moving this code to _get_params and returning empty dict will look a bit strange, IMO
There was a problem hiding this comment.
Well, the n is parameter, no?
There was a problem hiding this comment.
No, n is not a parameter, we only check if input is compatible with self.transformation_matrix and that's it. n wont change for the inputs.
There was a problem hiding this comment.
n is the product of the number of channels, height, and width. Since it
wont change for the inputs.
we should have it in _get_params with the
image = query_image(sample)
num_channels, height, width = F.get_image_dimensions(image)
n = num_channels * height * width
....
return dict(n=n)idiom.
There was a problem hiding this comment.
@vfdev-5 I think that's worth adding here to be consistent with other implementations. For tiny changes like this and the TODO below (#6458 (comment)) it might be worth doing them on the same PR. Otherwise we just increasing the TODO list which makes it harder to track and easier to miss.
| if inpt.device.type != self.mean_vector.device.type: | ||
| raise ValueError( | ||
| "Input tensor should be on the same device as transformation matrix and mean vector. " | ||
| f"Got {inpt.device} vs {self.mean_vector.device}" | ||
| ) |
There was a problem hiding this comment.
Given that this is easy to fix, should we maybe only warn and move mean and std to the same device?
There was a problem hiding this comment.
This is original implementation. Yes, we can move class attributes to provided device. Let's do that in a follow-up PR.
|
Hey @vfdev-5! You merged this PR, but no labels were added. The list of valid labels is available at https://github.com/pytorch/vision/blob/main/.github/process_commit.py |
| elif isinstance(inpt, PIL.Image.Image): | ||
| raise TypeError("Unsupported input type") |
There was a problem hiding this comment.
Idiomatic this should happen in forward with a has_any check.
|
@pmeier I do not agree with your |
Summary: * WIP * Fixed dtype correction and tests * Removed PIL Image support and output always Tensor Reviewed By: datumbox Differential Revision: D39013682 fbshipit-source-id: e98d9ba0f2ea703cfb9807dbc9abcd34e5db3331
Fixes #5538