Skip to content

1xDeJPG
Compare
Choose a tag to compare
@Phhofm Phhofm released this 24 Apr 15:23
· 39 commits to main since this release
1xDeJPG
c664eb5
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Learn about vigilant mode.

I thought I'll release 2 1xDeJPG models (SRFormerlight and OmniSR) I had been experimenting on. Maybe they are helpful.

Name: 1xDeJPG_SRFormer_light
pth(~36.8MB): https://drive.google.com/file/d/1qeoE-AymPBcuzItnjNsRrPLOw0chJN0F
onnx conversions in this folder which contains all files: https://drive.google.com/drive/folders/1RRXZCRVqlqU_iaeF-RGUM9UBWtI0HseY?usp=drive_link
Author: Philip Hofmann
Release Date: 02.08.2023
License: CC BY 4.0
Network: SRFormer light
Scale: 1
Purpose: 1xDeJPG model
Iterations: 88,000
batch_size: 4
HR_size: 32
Dataset: Nomos8k_sfw (processed with kim's datasetdestroyer - 2x jpg compression added)
Number of train images: 6118
OTF Training: No
Pretrained_Model_G: None

Description: 1xDeJPG model, meant to remove jpg artifacts. First compression round on the training dataset was 40-95, second round was 80-100.

Name: 1xDeJPG_OmniSR
pth(~6.2MB): https://drive.google.com/file/d/1y5_8pWm5hijzRiCYm5bsAg_N1kL5EWX
Author: Philip Hofmann
Release Date: 02.08.2023
License: CC BY 4.0
Network: OmniSR
Scale: 1
Purpose: A 1xDeJPG model
Iterations: 439,000
batch_size: 4
HR_size: 32
Dataset: Nomos8k_sfw (processed with kim's datasetdestroyer - 2x jpg compression added)
Number of train images: 6118
OTF Training: No
Pretrained_Model_G: None

Description: 1xDeJPG model, meant to remove jpg artifacts. First compression round on the training dataset was 40-95, second round was 80-100.

Name: 1xDeJPG_HAT
Author: Philip Hofmann
Release Date: 14.07.2023 (dd/mm/yy)
License: CC BY 4.0
Purpose: 1x dejpg model
Iterations: 54'000
batch_size: 4
HR_size: 32
Dataset: nomos8k_sfw
Number of train images: 2989
OTF Training: No
Pretrained_Model_G: None

Slowpoke Pics Example: https://slow.pics/c/70nGEvPR

PS these are non-otf models. otf experiments have not quite done it for me (probably would need more adjusting), but they can be found in the experiments folder and their outputs in the outputs folder: https://drive.google.com/drive/folders/1RRXZCRVqlqU_iaeF-RGUM9UBWtI0HseY?usp=drive_link

Assets 8
Loading