How to set image resolution for training and inference?

[sarmientoj24]

I cannot see it from the parameters.

Also, how to set which transforms to use?

[haoyuying]

For training, you can change the crop size,for example:https://github.com/PaddlePaddle/PaddleSeg/blob/release/2.6/configs/base/cityscapes_1024x1024.yml#L10 For inference, we take the original image without crop to predict the result.

[sarmientoj24]

that seems like random padding crop.

does that mean that the image should be in the expected size already? or is there a validation transforms in the yaml?

[github-actions[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed in 7 days if no further activity occurs. Thank you for your contributions.