Implement SwinIR v2 #2092
Merged
Implement SwinIR v2 #2092
+1,045
−7
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
Its nice that it kept the DOF , at least in this example |
|
That is interesting and clearly better. However I just want to point at the fact that there is still a lot of artefacts. Better, yes, definitely, but not so much IMHO. Thank you anyway. |
|
BSRGAN would be cool too |
|
wow its best technology |
6 tasks
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR implements SwinIR v2 from https://github.com/mv-lab/swin2sr. SwinIR v2 is SwinIR but with Swin Transformer V2 from Microsoft.
To use this, download
Swin2SR_RealworldSR_X4_64_BSRGAN_PSNR.pthfrom the aforementioned repo, and rename it to have the.v2.pthextension.The actual line addition of this PR is 28 lines, the rest is the model architecture.
Examples:
Original Image
SwinIR
SwinIR v2
In this example, SwinIR v2 avoids the blocking artifacts of v1, and also respects the (artistic?) blur/bokeh of the original image.