Info-icon.png We have moved to https://openmodeldb.info/. This new site has a tag and search system, which will make finding the right models for you much easier! If you have any questions, ask here: https://discord.gg/cpAUpDK

Difference between revisions of "Model Database"

From Upscale Wiki
Jump to navigation Jump to search
(Added the rest of the missing new models)
m (Updated link with more example images + changed model version to 1.1)
(32 intermediate revisions by 9 users not shown)
Line 6: Line 6:
 
=ESRGAN ("old Architecture") Models=
 
=ESRGAN ("old Architecture") Models=
  
Models that use the "old" ESRGAN architecture. They can be used either with the official ESRGAN repository (old arch tag) and BasicSR (old arch or preferably victorc's fork). You can also use tools that wrap around one of them, like IEU from Honh.
+
Models that use the "old" ESRGAN architecture. They can be used either with the official ESRGAN repository (old arch tag)
 +
and BasicSR (old arch or preferably victorc's fork). You can also use tools that wrap around one of them, like IEU from Honh.
  
 
==Image scaling and Video upscaling==
 
==Image scaling and Video upscaling==
Line 124: Line 125:
 
|10.242K
 
|10.242K
 
|Pokemon DVD - BluRay
 
|Pokemon DVD - BluRay
 +
|[https://mega.nz/#!vtgSWKQT!K7Asn2zKe4N70R2aV89KEMTKhH3aiyGAAiuQDJF09qs 2xESRGAN]
 +
|-
 +
|[https://www115.zippyshare.com/v/O2do1VCf/file.html Pokemodel Lite]
 +
|Neo-Raws#4055
 +
|2
 +
|[http://www.wtfpl.net/ WTFPL]
 +
|Old Anime
 +
|100K
 +
|4
 +
|128
 +
|268
 +
|Pokemon
 
|[https://mega.nz/#!vtgSWKQT!K7Asn2zKe4N70R2aV89KEMTKhH3aiyGAAiuQDJF09qs 2xESRGAN]
 
|[https://mega.nz/#!vtgSWKQT!K7Asn2zKe4N70R2aV89KEMTKhH3aiyGAAiuQDJF09qs 2xESRGAN]
 
|-
 
|-
Line 185: Line 198:
 
|Misc digital art, material studies
 
|Misc digital art, material studies
 
|4xfalcoon300(manga)
 
|4xfalcoon300(manga)
 +
|-
 +
|[https://drive.google.com/file/d/1FWzEgqaSg2z0qOSjcmAxEUB3clEkm1PG/view?usp=sharing 2x_Byousoku_5_Centimeter]
 +
|Mystery_Bullet#0642 (Shady Adel)
 +
|2
 +
|[https://www.gnu.org/licenses/gpl-3.0.en.html GNU GPLv3]
 +
|Anime landscape upscale
 +
|300k
 +
|8
 +
|128
 +
|3K
 +
|Frames from BluRay of Byousoku 5 Centimeter
 +
|[https://mega.nz/#!vtgSWKQT!K7Asn2zKe4N70R2aV89KEMTKhH3aiyGAAiuQDJF09qs 2xESRGAN]
 
|-
 
|-
 
|[https://u.pcloud.link/publink/show?code=kZb2mFXZOwIwPolF8HYAlU8QE04StpEC9YRX UltraYandere]
 
|[https://u.pcloud.link/publink/show?code=kZb2mFXZOwIwPolF8HYAlU8QE04StpEC9YRX UltraYandere]
Line 197: Line 222:
 
|Yandere4800 (yande.re artworks)
 
|Yandere4800 (yande.re artworks)
 
|[https://mega.nz/#!uhAVEaAA!qfwIQ44Ba3rXMAb2-M3aM3zFBCwzmyQ64IO_0O-csJE 4xESRGAN]
 
|[https://mega.nz/#!uhAVEaAA!qfwIQ44Ba3rXMAb2-M3aM3zFBCwzmyQ64IO_0O-csJE 4xESRGAN]
 +
|-
 +
|[https://u.pcloud.link/publink/show?code=kZcyv0XZOefT8dS3Vu0fsPXV6U9ELQL2CO6X PatchySharp]
 +
|[[User:Nmkd|nmkd]]
 +
|4
 +
|[http://www.wtfpl.net/ WTFPL]
 +
|Upscaler for clean images or images with compression artifacts
 +
|240k
 +
|5
 +
|256
 +
|1800
 +
|DIV2K Clean + JPEG
 +
||[https://mega.nz/#!uhAVEaAA!qfwIQ44Ba3rXMAb2-M3aM3zFBCwzmyQ64IO_0O-csJE 4xESRGAN]
 
|}
 
|}
  
 
Manga109Attempt is slightly blurry, but performs well as a general upscaler.
 
Manga109Attempt is slightly blurry, but performs well as a general upscaler.
  
Falcoon Fanart tries to improve upon it with the goal of removing checkerboard patterns / and dithering. It has oil color based shading with sharp lines.
+
Falcoon Fanart tries to improve upon it with the goal of removing checkerboard patterns / and dithering. It has oil color
 +
based shading with sharp lines.
  
WaifuGAN v3 is [[User:DinJerr|DinJerr]]’s third attempt at training from a mostly anime dataset sourced from image boards and is intended for upscaling CG-painted anime with variable outlines. Only PNGs were used, mainly with brush strokes and gradients. Texturised images avoided as much as possible. If too generative, tone down by interpolating with a softer model.
+
WaifuGAN v3 is [[User:DinJerr|DinJerr]]’s third attempt at training from a mostly anime dataset sourced from image boards
 +
and is intended for upscaling CG-painted anime with variable outlines. Only PNGs were used, mainly with brush strokes and
 +
gradients. Texturised images avoided as much as possible. If too generative, tone down by interpolating with a softer model.
  
De-Toon, is a model that does the opposite of tooning an image. It takes toon style shading and detail, and attempts to make it realistic. Its very sensitive, and can be used on small sprites, to large images. Also included is a alt version, which is less sharp.
+
De-Toon, is a model that does the opposite of tooning an image. It takes toon style shading and detail, and attempts to
 +
make it realistic. Its very sensitive, and can be used on small sprites, to large images. Also included is a alt version,
 +
which is less sharp.
  
 
DigiPaint is a digital art upscaler designed to take brush strokes into account, based off of the Falcoon Fanart model.
 
DigiPaint is a digital art upscaler designed to take brush strokes into account, based off of the Falcoon Fanart model.
  
DeviantPixelHD was designed to upscale original LucasArts games' backgrounds. High Definition digital art from Deviant Art was used, reducing the LR training to 32 colors so the training images looked pixelated. Can be used with other environments such as manga or digital art. Examples can be found on Raúl Sangonzalo's YouTube account (<nowiki>https://www.youtube.com/channel/UCwBfuiHdSPQ-zslOmiW8OHg</nowiki>).
+
DeviantPixelHD was designed to upscale original LucasArts games' backgrounds. High Definition digital art from Deviant
 +
Art was used, reducing the LR training to 32 colors so the training images looked pixelated. Can be used with other
 +
environments such as manga or digital art. Examples can be found on Raúl Sangonzalo's YouTube account
 +
(<nowiki>https://www.youtube.com/channel/UCwBfuiHdSPQ-zslOmiW8OHg</nowiki>).
  
 
FireAlpha is an attempt to create a 4 channel model that handles transparency without splitting and upscaling separately.
 
FireAlpha is an attempt to create a 4 channel model that handles transparency without splitting and upscaling separately.
Line 215: Line 260:
 
*'''''Notice:''' IEU can use the model but will attempt to split and handle transparency itself, BlueAmulet's fork of ESRGAN is required to use this model. https://github.com/BlueAmulet/ESRGAN''
 
*'''''Notice:''' IEU can use the model but will attempt to split and handle transparency itself, BlueAmulet's fork of ESRGAN is required to use this model. https://github.com/BlueAmulet/ESRGAN''
  
4xFSDedither_Manga was intended for pixel art containing dithering patterns, but ended up being effective on pixel art and manga/cartoon style images without any dithering at all.
+
4xFSDedither_Manga was intended for pixel art containing dithering patterns, but ended up being effective on pixel art
 +
and manga/cartoon style images without any dithering at all.
  
 
====Cartoon / Comic====
 
====Cartoon / Comic====
Line 257: Line 303:
 
|SD to HD film rescan
 
|SD to HD film rescan
 
| Scratch
 
| Scratch
 +
|-
 +
|[https://drive.google.com/file/d/1G3mTGd6lF7SgfKyb-hdkfD-FfMHCWk_h/view?usp=sharing CelFrames 1.0]
 +
|Klexos
 +
|4
 +
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 +
|Film Animation
 +
|20K
 +
|38
 +
|128
 +
|30k
 +
|4K Animation Cel Scans
 +
| DigitalFrames 2.0
 
|-
 
|-
 
|[https://mega.nz/#!twxxWQTT!OjPL0o9sq2MM6Flac1shLo9CvE33P9mKKF73OMT-j2I DigitalFrames 2.0]
 
|[https://mega.nz/#!twxxWQTT!OjPL0o9sq2MM6Flac1shLo9CvE33P9mKKF73OMT-j2I DigitalFrames 2.0]
Line 282: Line 340:
 
|DigitalFrames 2.0
 
|DigitalFrames 2.0
 
|-
 
|-
|[https://drive.google.com/file/d/1RhEkTYy-1DxJP2ynLW5NSrz6BjVMy8xx/view Struzan]
+
|-
 +
|[https://drive.google.com/file/d/1XdJdzcCQy7eV98bOEk9S5CfOrjBbSnZW/view?usp=sharing DigitalFrames 2.5]
 +
|Klexos
 +
|4
 +
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 +
|Digital Animation
 +
|925K
 +
|38
 +
|128
 +
|30K
 +
|SD to HD Cartoons
 +
| DigitalFrames 2.0
 +
|-
 +
|[https://drive.google.com/file/d/1RhEkTYy-1DxJP2ynLW5NSrz6BjVMy8xx/view Struzan]
 
|Laserschwert
 
|Laserschwert
 
|4
 
|4
Line 389: Line 460:
 
|Variety of images from ArtStation
 
|Variety of images from ArtStation
 
|HugePeeps
 
|HugePeeps
 +
|-
 +
|[http://cloud.mitch.best/index.php/s/iio9SL9BdgWiCoi 4x_AbeScale]
 +
|mitch#1337
 +
|4
 +
|GNU
 +
|Linework Cartoons
 +
|110k
 +
|8
 +
|192
 +
|8
 +
|LR: Clone High DVD, HR: Re-illustrated Vectors from Frames
 +
|4xESRGAN
 
|}
 
|}
  
The Comic Book model was trained using stills from the film spiderman into the spiderverse, has a comic book crosshatch shading effect to the images. [https://cdn.discordapp.com/attachments/547950274313191425/564427623204388864/6de7c88f.png Sample]
+
The Comic Book model was trained using stills from the film spiderman into the spiderverse, has a comic book crosshatch
 +
shading effect to the images. [https://cdn.discordapp.com/attachments/547950274313191425/564427623204388864/6de7c88f.png Sample]
  
FilmFrames 0.5: Purpose of the model is upscaling medium quality sourced SD animation frames to an HD "Film rescan" status. https://imgsli.com/MjA2NDM
+
FilmFrames 0.5: Purpose of the model is upscaling medium quality sourced SD animation frames to an HD "Film rescan" status.
 +
https://imgsli.com/MjA2NDM
  
DeteriorationFrames 1.0: The purpose of this model is to aid in the creation of datasets aimed at animation by saving you time and allowing you to deteriorate the HQ/HD frames directly though this model instead of having to go through frame by frame to find SD/LQ counterparts. This iteration in particular adds dot crawl, ringing and chromatic aberrations. https://imgsli.com/MTk4MjA.
+
CelFrames 1.0: Trained on thousands of 4K high quality cel scans, the purpose of this model is restore the frame to it's original Cel look, with a particular emphasys on textures and preserving those brushstrokes and natural lines. '''Not to be used with heavily grainy sources.''' https://imgsli.com/Mzk2NzU
 +
 
 +
DeteriorationFrames 1.0: The purpose of this model is to aid in the creation of datasets aimed at animation by saving you time
 +
and allowing you to deteriorate the HQ/HD frames directly though this model instead of having to go through frame by frame to
 +
find SD/LQ counterparts. This iteration in particular adds dot crawl, ringing and chromatic aberrations. https://imgsli.com/MTk4MjA.
 
Important note: Resizing your HD counterparts before using this model to create the LRs is essential.
 
Important note: Resizing your HD counterparts before using this model to create the LRs is essential.
Tip: If you blur the HRs after resizing them but before using the deterioration model on them, depending on how much you blur you'll get less Chromatic aberrations. So with this you can tweak both blurriness and the aberrations to match better the source you intend to actually use the model you are about to train on.
+
Tip: If you blur the HRs after resizing them but before using the deterioration model on them, depending on how much you blur
 +
you'll get less Chromatic aberrations. So with this you can tweak both blurriness and the aberrations to match better the source
 +
you intend to actually use the model you are about to train on.
  
DigitalFrames 2.0: Was trained completely separately from 1.0 with a completely new and bigger dataset. It has a reduced sensitivity to noise which reduces the blobs of noise phenomena. It can also work reasonably well on Cel/Film Toons now. Overall it's a big leap from 1.0.
+
DigitalFrames 2.0: Was trained completely separately from 1.0 with a completely new and bigger dataset. It has a reduced sensitivity
 +
to noise which reduces the blobs of noise phenomena. It can also work reasonably well on Cel/Film Toons now. Overall it's a big leap
 +
from 1.0.
  
DigitalFrames 2.1_Aggressive: Provides the sharpest results yet. It's the most aggressive version of DF. Does it's best on animation with little to no textures. The model as well as any version of DF is not suitable for all kinds of animation.
+
DigitalFrames 2.1_Aggressive: Provides the sharpest results yet. It's the most aggressive version of DF. Does it's best on animation
 +
with little to no textures. The model as well as any version of DF is not suitable for all kinds of animation.
 +
 
 +
DigitalFrames 2.5: Big leap from 2.0, especially as much as texture retention is concerned https://imgsli.com/Mzk2MzU.
  
 
FatalimiX is specialized in making those low res comics high res again!
 
FatalimiX is specialized in making those low res comics high res again!
  
The SpongeBob model was trained for the purpose of upscaling Battle for Bikini Bottom's textures, but it works well for upscaling scenes from the show as well. It was initially trained on the clean downscaled images but later JPEG, quantization, dithering, and Gaussian blur were added to the LR OTF options to help it upscale the textures better.
+
The SpongeBob model was trained for the purpose of upscaling Battle for Bikini Bottom's textures, but it works well for upscaling
 +
scenes from the show as well. It was initially trained on the clean downscaled images but later JPEG, quantization, dithering, and
 +
Gaussian blur were added to the LR OTF options to help it upscale the textures better.
  
Spongebob v6 is ideally a lot sharper and cleaner, but Joey's still not sure if it works better than the old one. It still seems to be better in many cases.
+
Spongebob v6 is ideally a lot sharper and cleaner, but Joey's still not sure if it works better than the old one. It still seems to
 +
be better in many cases.
  
 
Spongebob v6 De-Quantize is for de-quantizing as well as upscaling. The results are pretty blurry but it works decently enough.
 
Spongebob v6 De-Quantize is for de-quantizing as well as upscaling. The results are pretty blurry but it works decently enough.
  
Spongebob v6 Deblur was created as Joey wasn't entirely happy with the main Spongebob v6 model. It was trained with blurring OTF options and two different downscale types.
+
Spongebob v6 Deblur was created as Joey wasn't entirely happy with the main Spongebob v6 model. It was trained with blurring OTF
 +
options and two different downscale types.
  
HugePeeps v1 was created by DinJerr to test out a theory on how to properly train 8x models. This particular model best works on painted pics of ladies.
+
HugePeeps v1 was created by DinJerr to test out a theory on how to properly train 8x models. This particular model best works on
 +
painted pics of ladies.
  
 
HugePaint v1 is trained on a more generic dataset with a balance of illustrated humans, vegetation, minerals and some animals.
 
HugePaint v1 is trained on a more generic dataset with a balance of illustrated humans, vegetation, minerals and some animals.
  
Some of the first cartoon painted models for ESRGAN can still be found [https://esrgan.blogspot.com/2019/01 here], but have been mostly superseded by newer models.
+
Abescale is an upscale model intended to bring Clone High DVDs (480i/p) to 4k. Based on re-illustrated frames from the DVD, is effective on similar cartoons with heavy linework.
 +
 
 +
Some of the first cartoon painted models for ESRGAN can still be found [https://esrgan.blogspot.com/2019/01 here], but have been
 +
mostly superseded by newer models.
  
 
====Pixel Art====
 
====Pixel Art====
  
Pixel art is a form of digital art, created through the use of software, where images are edited on the pixel level. The aesthetic for this kind of graphics comes from 8-bit and 16-bit computers and video game consoles, in addition to other limited systems such as graphing calculators. In most pixel art, the color palette used is extremely limited in size, with some pixel art using only two colors.
+
used in tandem with other bitmaps to construct a larger scene.
 
 
Creating or modifying pixel art characters or objects for video games is sometimes called spriting, a term that arose from the hobbyist community. The term likely came from the term sprite, a term used in computer graphics to describe a two-dimensional bitmap that is used in tandem with other bitmaps to construct a larger scene.
 
 
{| class="wikitable sortable"
 
{| class="wikitable sortable"
 
|-
 
|-
Line 576: Line 678:
 
|4
 
|4
 
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
|(dithered) spirites
+
|(dithered) sprites
 
|265K
 
|265K
 
|10
 
|10
Line 582: Line 684:
 
|19.7K
 
|19.7K
 
|?
 
|?
|[https://de-next.owncube.com/index.php/s/YbAqNMMTtxrajF6 Face]
+
|link
 +
|-
 +
|[https://u.pcloud.link/publink/show?code=kZ4WthXZDDnqbo4Hye8zBgCkhgbbQf3WXT27 Arzenal (v1.1)]
 +
|[[User:Glasshopper|Glasshopper]]
 +
|8
 +
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 +
|Smooth general pixel art, Minecraft textures
 +
|175K
 +
|?
 +
|?
 +
|?
 +
|None used, model was interpolated
 +
|[https://nmkd.de/?esrgan-list nmkd's pixel art upscaling models with different interpolation settings and de-dithered versions]
 
|-
 
|-
|[https://de-next.owncube.com/index.php/s/ECsEHxdoYCnFsZA Fatal Pixels]
+
 
 +
|[https://de-next.owncube.com/index.php/s/NszBMb7T5wYG9X6 Fatality MK2]
 
|[[User:Twittman|Twittman]]
 
|[[User:Twittman|Twittman]]
 
|4
 
|4
 
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 
|Sprites
 
|Sprites
|340K
+
|90K
 +
|10
 +
|128
 +
|42K
 +
|Anime, Manga and some real life
 +
|A previously attempted MK2
 +
|-
 +
|[https://de-next.owncube.com/index.php/s/ECsEHxdoYCnFsZA Fatal Pixels]
 +
|[[User:Twittman|Twittman]]
 +
|4
 +
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 +
|Sprites
 +
|340K
 
|22
 
|22
 
|96
 
|96
Line 595: Line 722:
 
|Anime, Manga
 
|Anime, Manga
 
|Fatality_MKII_90k
 
|Fatality_MKII_90k
 +
|-
 +
|[https://drive.google.com/drive/u/1/folders/1C82o6p9jXzOZIZtMHeVXFW8BYV0s10FA PixelPerfect]
 +
|[[User:Mutin Choler|Mutin Choler]]
 +
|4
 +
|[http://www.wtfpl.net/ WTFPL]
 +
|Sprites
 +
|137K
 +
|4
 +
|128
 +
|300
 +
|Anime, Manga
 +
|1xESRGAN
 
|}
 
|}
  
Fatality is meant to be used for upscaling medium resolution Sprites, dithered or un-dithered, it can also upscale manga/anime and gameboy camera images.
+
Fatality is meant to be used for upscaling medium resolution Sprites, dithered or un-dithered, it can also upscale manga/anime
 +
and gameboy camera images.
 +
 
 +
Arzenal is a mashup of multiple of nmkd's pixel art models, including de-dithered version for noise reduction, specifically designed to work on many different kinds of pixel art, without being too smooth, or too rough. The reason being is that nmkd's models are good, but they are TOO good for their purposes, and a model that evens them out can create results that fit both tasks. The de-dithering was also added for smoothing, and noise reduction.
  
Rebout is trained to give detail to character models, with faces and hands improved. Based on the snk game kof94 rebout, although best for snk style games, does work on a variety of sprites. Also included is a interpolated version that may provide a cleaner upscale for certain sprites.
+
Rebout is trained to give detail to character models, with faces and hands improved. Based on the snk game kof94 rebout, although
 +
best for snk style games, does work on a variety of sprites. Also included is a interpolated version that may provide a cleaner
 +
upscale for certain sprites.
  
Faithful 2x was trained on the faithful 32x32 Minecraft texture/resource pack, using both the 1.13 and 1.14 versions (which have different textures). For those not familiar, this texture pack is meant to upscale the 16x16 textures to 32x32 faithfully, as in make them as close as possible to the original, but higher resolution. The model does a decent job at replicating this.
+
Faithful 2x was trained on the faithful 32x32 Minecraft texture/resource pack, using both the 1.13 and 1.14 versions (which have
 +
different textures). For those not familiar, this texture pack is meant to upscale the 16x16 textures to 32x32 faithfully, as in
 +
make them as close as possible to the original, but higher resolution. The model does a decent job at replicating this.
  
*'''NOTE:''' Due to an error in the dataset, sometimes the model turns jagged white lines green. Only encountered this in one texture tried, so it appears unlikely to happen.
+
*'''NOTE:''' Due to an error in the dataset, sometimes the model turns jagged white lines green. Only encountered this in one texture
 +
tried, so it appears unlikely to happen.
  
FArtDIV3 Suite is a group of models based on a large image dataset trained using OTF tile generation. Most of the models are based off the Base model, and suite different purposes. Fine is used to generate details, but may be noisy/grainy at times. Blend will attempt to blend colour steps into proper gradients. UnholyFArt is a interpolated model that reduces details in favor of sharp blends.
+
FArtDIV3 Suite is a group of models based on a large image dataset trained using OTF tile generation. Most of the models are based
 +
off the Base model, and suite different purposes. Fine is used to generate details, but may be noisy/grainy at times. Blend will
 +
attempt to blend colour steps into proper gradients. UnholyFArt is a interpolated model that reduces details in favor of sharp blends.
  
 
BigFArt Suite is similar to FArtDIV3, but for larger pixel sprites/art in the context of tile size.
 
BigFArt Suite is similar to FArtDIV3, but for larger pixel sprites/art in the context of tile size.
  
BigFace_v3 Suite takes the BigFace model from the BigFArt suite, and tries to improve by using double-sized tiles. Main model has good dedithering. _Blend focuses on sharp blends, and _Clear will abstract and sharpen details.
+
BigFace_v3 Suite takes the BigFace model from the BigFArt suite, and tries to improve by using double-sized tiles. Main model has good
 +
dedithering. _Blend focuses on sharp blends, and _Clear will abstract and sharpen details.
  
 
===Photographs and Photorealism===
 
===Photographs and Photorealism===
 
A photograph (also known as a photo) is an image created by light falling on a photosensitive surface, usually photographic film or an electronic image sensor, such as a CCD or a CMOS chip. Most photographs are created using a camera, which uses a lens to focus the scene's visible wavelengths of light into a reproduction of what the human eye would see. The process and practice of creating such images is called photography. The word photograph was coined in 1839 by Sir John Herschel and is based on the Greek φῶς (phos), meaning "light," and γραφή (graphê), meaning "drawing, writing," together meaning "drawing with light."
 
Photorealism is a genre of art that encompasses painting, drawing and other graphic media, in which an artist studies a photograph and then attempts to reproduce the image as realistically as possible in another medium. Although the term can be used broadly to describe artworks in many different media, it is also used to refer specifically to a group of paintings and painters of the American art movement that began in the late 1960s and early 1970s.
 
  
 
====Misc / Kitchen Sink====
 
====Misc / Kitchen Sink====
  
 
All kinds of photographs or photorealistic images. Those models aren't specialized.
 
All kinds of photographs or photorealistic images. Those models aren't specialized.
 +
 
{| class="wikitable sortable"
 
{| class="wikitable sortable"
 
|-
 
|-
Line 689: Line 837:
 
|Universal
 
|Universal
 
|190K
 
|190K
 +
|4
 +
|128
 +
|1400K
 +
|DIV2K + Anime
 +
|[https://mega.nz/#!uhAVEaAA!qfwIQ44Ba3rXMAb2-M3aM3zFBCwzmyQ64IO_0O-csJE 4xESRGAN]
 +
|-
 +
|[https://drive.google.com/file/d/1rRT27g6id9eOz-12wb1xjOnVSEDOOMdx/view?usp=sharing UniversalUpscaler-Detailed]
 +
|[[User:Mutin Choler|Mutin Choler]]
 +
|4
 +
|WTFPL
 +
|Universal
 +
|155K
 
|4
 
|4
 
|128
 
|128
Line 720: Line 880:
 
|}
 
|}
  
Box was meant to be an improvement on the RRDB_ESRGAN_x4 model ([https://imgsli.com/NDEwMw comparison]). It’s also trained on photos, but with a much larger dataset which was downscaled with linear interpolation (box filter) instead of bicubic.
+
Box was meant to be an improvement on the RRDB_ESRGAN_x4 model ([https://imgsli.com/NDEwMw comparison]). It’s also trained
 +
on photos, but with a much larger dataset which was downscaled with linear interpolation (box filter) instead of bicubic.
  
Deoldify is intended for denoising and deblurring photos. All outputs are greyscale. This was initially a proof of concept. ''(Need clarification on specific subject matter.)''
+
Deoldify is intended for denoising and deblurring photos. All outputs are greyscale. This was initially a proof of concept.
 +
''(Need clarification on specific subject matter.)''
  
 
The Ground model was trained on various pictures of '''stones, dirt and grass''' using Google’s image search.
 
The Ground model was trained on various pictures of '''stones, dirt and grass''' using Google’s image search.
  
The Misc model is trained on various pictures shoot by myself, including '''bricks, stone, dirt, grass, plants, wood, bark, metal''' and a few others.
+
The Misc model is trained on various pictures shoot by myself, including '''bricks, stone, dirt, grass, plants, wood, bark, metal'''
 +
and a few others.
  
Superscale by NMKD was made to rival the RealSR models and is one of the best upscalers for real-world content. It's trained on PNGs, so it doesn't handle compression artifacts well.
+
Superscale by NMKD was made to rival the RealSR models and is one of the best upscalers for real-world content. It's trained on PNGs,
 +
so it doesn't handle compression artifacts well.
  
Lollypop was originally made for prerendered manga faces, using the patchgan discriminator. Combined with cutmixup, cx loss, and Frequency Separator, it handles prerendered, realistic faces, and even pixel art, with dedithering low palette images incredibly well.
+
Lollypop was originally made for prerendered manga faces, using the patchgan discriminator. Combined with cutmixup, cx loss, and
 +
Frequency Separator, it handles prerendered, realistic faces, and even pixel art, with dedithering low palette images incredibly well.
  
 
====Characters and Faces====
 
====Characters and Faces====
Line 786: Line 951:
 
|Custom (Faces)
 
|Custom (Faces)
 
|4x_Faces_04_N_180000_G.pth
 
|4x_Faces_04_N_180000_G.pth
 +
|-
 +
|[https://f002.backblazeb2.com/file/ESRGAN/_RELEASE/4x_cat_patch_325000_G.pth Cat_Patch]
 +
|[[User:Twittman|Twittman]]
 +
|4
 +
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 +
|For upscaling Meow
 +
|325k
 +
|8
 +
|128
 +
|9.5k
 +
|Custom (Cats)
 +
|previous attempt
 
|-
 
|-
 
|[https://drive.google.com/open?id=1OyOJIW224hBhb-aTCbuUQb0qzKmE4oH6 TGHQFace8x]
 
|[https://drive.google.com/open?id=1OyOJIW224hBhb-aTCbuUQb0qzKmE4oH6 TGHQFace8x]
Line 800: Line 977:
 
|}
 
|}
  
Trixie was made to bring balance to the force… Also to upscale '''character textures for star wars games''', including the heroes, rebels, sith and imperial. Plus a few main aliens…Why called trixie? Because jar jars big adventure would be too long of a name…This also provides good upscale for '''face textures for general purpose as well as basic star wars''' textures.
+
Trixie was made to bring balance to the force… Also to upscale '''character textures for star wars games''', including the heroes,
 +
rebels, sith and imperial. Plus a few main aliens…Why called trixie? Because jar jars big adventure would be too long of a name…
 +
This also provides good upscale for '''face textures for general purpose as well as basic star wars''' textures.
  
The Face Focus modes was designed for '''slightly out of focus / blurred images of faces'''. It is aimed at '''faces''' / '''hair''', but it can help to '''improve other out of focused images too''' as always just try it.
+
The Face Focus modes was designed for '''slightly out of focus / blurred images of faces'''. It is aimed at '''faces''' / '''hair''',
 +
but it can help to '''improve other out of focused images too''' as always just try it.
  
 
===Specialized===
 
===Specialized===
Line 833: Line 1,013:
 
|1.176k
 
|1.176k
 
|Images of Coins
 
|Images of Coins
 +
|[https://mega.nz/#!uhAVEaAA!qfwIQ44Ba3rXMAb2-M3aM3zFBCwzmyQ64IO_0O-csJE 4xESRGAN]
 +
|-
 +
|[https://drive.google.com/file/d/1JI8VZSSHpH9u7HzE8MvuNG351Dgn4jwU/view BS_Screenbooster_SPSR]
 +
|[[User:BlackScout|BlackScout]]
 +
|4
 +
|[https://www.gnu.org/licenses/gpl-3.0.en.html GNU GPLv3]
 +
|Upscale game screenshots
 +
|76K
 +
|2
 +
|128
 +
|1.2M (tiles)
 +
|Game screenshots
 
|[https://mega.nz/#!uhAVEaAA!qfwIQ44Ba3rXMAb2-M3aM3zFBCwzmyQ64IO_0O-csJE 4xESRGAN]
 
|[https://mega.nz/#!uhAVEaAA!qfwIQ44Ba3rXMAb2-M3aM3zFBCwzmyQ64IO_0O-csJE 4xESRGAN]
 
|-
 
|-
Line 944: Line 1,136:
 
|}
 
|}
  
The map model was trained on '''maps, old documents, papers and various styles of typefaces/fonts'''. Based on a dataset contributed by alsa. [https://imgsli.com/MzMwNw Sample]
+
The map model was trained on '''maps, old documents, papers and various styles of typefaces/fonts'''. Based on a dataset contributed
 +
by alsa. [https://imgsli.com/MzMwNw Sample]
  
 
The Forest model is focused on '''trees, leaves, bark and stone''' can be used for double upscaling for even more detail. [https://imgsli.com/MzI2Nw Sample]
 
The Forest model is focused on '''trees, leaves, bark and stone''' can be used for double upscaling for even more detail. [https://imgsli.com/MzI2Nw Sample]
  
The Armory model was trained with '''modded textures form Skyrim''', including '''Clothing, Armor and Weapons'''. ('''Leather, Canvas and Metal''' should all work - maybe too sharp so interpolate)
+
The Armory model was trained with '''modded textures form Skyrim''', including '''Clothing, Armor and Weapons'''. ('''Leather, Canvas and Metal'''
 +
should all work - maybe too sharp so interpolate)
  
 
The wood model was trained for Skyrim by Laeris.
 
The wood model was trained for Skyrim by Laeris.
  
The Skyrim Diffuse models is supposed to be used with Skyrim’s diffuse textures. It is a bit too sharp so I recommend to interpolating with the |[https://mega.nz/#!uhAVEaAA!qfwIQ44Ba3rXMAb2-M3aM3zFBCwzmyQ64IO_0O-csJE 4xESRGAN] model or the mangaAttempt109 model, look in [https://drive.google.com/drive/folders/1m1sgHiSZkmI4NbiXeBEZasf94WPqEsNF Deorder’s Skyrim Model Google Drive] for an already interpolated version.
+
The Skyrim Diffuse models is supposed to be used with Skyrim’s diffuse textures. It is a bit too sharp so I recommend to interpolating
 +
with the |[https://mega.nz/#!uhAVEaAA!qfwIQ44Ba3rXMAb2-M3aM3zFBCwzmyQ64IO_0O-csJE 4xESRGAN] model or the mangaAttempt109 model, look in
 +
[https://drive.google.com/drive/folders/1m1sgHiSZkmI4NbiXeBEZasf94WPqEsNF Deorder’s Skyrim Model Google Drive] for an already interpolated version.
  
 
Fallout 4 weapons was trained using Fallout 4’s official HDarmor/weapon textures but could be used on other weapon and armor textures.
 
Fallout 4 weapons was trained using Fallout 4’s official HDarmor/weapon textures but could be used on other weapon and armor textures.
Line 958: Line 1,154:
 
===Grayscale===
 
===Grayscale===
  
In digital photography, computer-generated imagery, and colorimetry, a grayscale or greyscale image is one in which the value of each pixel is a single sample representing only an amount of light, that is, it carries only intensity information. Grayscale images, a kind of black-and-white or gray monochrome, are composed exclusively of shades of gray. The contrast ranges from black at the weakest intensity to white at the strongest.[1]
 
 
{| class="wikitable sortable"
 
{| class="wikitable sortable"
 
|-
 
|-
Line 1,057: Line 1,252:
 
|[https://drive.google.com/open?id=1f5kIr3gG5GjbN_EM96CCB0nqu_2_56gh ISO denoise v1]
 
|[https://drive.google.com/open?id=1f5kIr3gG5GjbN_EM96CCB0nqu_2_56gh ISO denoise v1]
 
|-
 
|-
|[https://drive.google.com/file/d/1hnWvQ2VzzxaTaar6ifErFJKrCmYXH8v0/view?usp=sharing NoiseToner]
+
|[https://drive.google.com/drive/u/1/folders/12j6BvuR7SU48ui9DHjBPB1vzPpQri9bD NoiseToner (Uniform)]
 
|[[User:Mutin Choler|Mutin Choler]]
 
|[[User:Mutin Choler|Mutin Choler]]
 
|1
 
|1
 
|[http://www.wtfpl.net/ WTFPL]
 
|[http://www.wtfpl.net/ WTFPL]
|Remove high ISO noise
+
|Remove Uniform noise
|105K
 
|4
 
|64
 
|800
 
|Custom (Photos)
 
|[https://mega.nz/#!D940CAKR!-BAx7tSj1CgeopGkaMW31VuPsC3hvNBL35TEVAj4LIo 1xESRGAN]
 
|-
 
|[https://drive.google.com/file/d/1hnWvQ2VzzxaTaar6ifErFJKrCmYXH8v0/view?usp=sharing NoiseToner Uniform]
 
|[[User:Mutin Choler|Mutin Choler]]
 
|1
 
|[http://www.wtfpl.net/ WTFPL]
 
|Remove noise
 
 
|100K
 
|100K
 
|4
 
|4
Line 1,081: Line 1,264:
 
|[https://mega.nz/#!D940CAKR!-BAx7tSj1CgeopGkaMW31VuPsC3hvNBL35TEVAj4LIo 1xESRGAN]
 
|[https://mega.nz/#!D940CAKR!-BAx7tSj1CgeopGkaMW31VuPsC3hvNBL35TEVAj4LIo 1xESRGAN]
 
|-
 
|-
|[https://drive.google.com/file/d/1hnWvQ2VzzxaTaar6ifErFJKrCmYXH8v0/view?usp=sharing NoiseToner Poisson]
+
|[https://drive.google.com/drive/u/1/folders/12j6BvuR7SU48ui9DHjBPB1vzPpQri9bD NoiseToner (Poisson)]
 
|[[User:Mutin Choler|Mutin Choler]]
 
|[[User:Mutin Choler|Mutin Choler]]
 
|1
 
|1
Line 1,093: Line 1,276:
 
|[https://mega.nz/#!D940CAKR!-BAx7tSj1CgeopGkaMW31VuPsC3hvNBL35TEVAj4LIo 1xESRGAN]
 
|[https://mega.nz/#!D940CAKR!-BAx7tSj1CgeopGkaMW31VuPsC3hvNBL35TEVAj4LIo 1xESRGAN]
 
|}
 
|}
High ISO is required to capture images at very high shutter speeds and/or in poor lighting. However, increasing ISO also results in noise that is challenging to remove without also losing fine details. To remove this noise without damaging details, one can apply the technique of median stacking: combine 10-100 photos of the exact same object to separate the signal (clear photo) from noise.
+
High ISO is required to capture images at very high shutter speeds and/or in poor lighting. However,
 +
increasing ISO also results in noise that is challenging to remove without also losing fine details.
 +
To remove this noise without damaging details, one can apply the technique of median stacking: combine
 +
10-100 photos of the exact same object to separate the signal (clear photo) from noise.
  
ISO denoise v1 is trained on a handcraft set of pairs between original noisy photos and median stacked counterparts. It tends to be conservative, avoiding modifying anything that does not closely resemble camera ISO noise. ISO denoise v2 uses a larger dataset, handling some cases better and some cases worse than v1. If results are unsatisfactory with one model, try the other model.
+
ISO denoise v1 is trained on a handcraft set of pairs between original noisy photos and median stacked
 +
counterparts. It tends to be conservative, avoiding modifying anything that does not closely resemble
 +
camera ISO noise. ISO denoise v2 uses a larger dataset, handling some cases better and some cases worse
 +
than v1. If results are unsatisfactory with one model, try the other model.
  
 
===JPG Compression===
 
===JPG Compression===
Line 1,294: Line 1,483:
 
|Real life, Manga and digital text
 
|Real life, Manga and digital text
 
|1x_DeJpeg_Fatality_01_200000_G.pth
 
|1x_DeJpeg_Fatality_01_200000_G.pth
 +
|-
 +
|[https://f002.backblazeb2.com/file/ESRGAN/_RELEASE/1x_Saiyajin_DeJPEG_300000_G.pth SaiyaJin DeJpeg]
 +
|[[User:Twittman|Twittman]]
 +
|1
 +
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 +
|JPG compressed Images from 2% to 100% quality
 +
|300K
 +
|1/8
 +
|128
 +
|22.8K
 +
|Anime, real life, manga, generated
 +
|[https://de-next.owncube.com/index.php/s/w82HLrLWmWi4SQ5 1x_DeJpeg_Fatality_PlusULTRA_200000_G.pth]
 
|}
 
|}
  
JPG gets compressed with a Quality Percentage between 0 and 100. So depending on how bad your JPEGs are compressed, choose the model of your choice. You can use [https://imagemagick.org/index.php ImageMagick] to guess the Quality percentage, but keep in mind that it might be wrong, since the image might have been re-saved.
+
JPG gets compressed with a Quality Percentage between 0 and 100. So depending on how bad your JPEGs are compressed, choose
 +
the model of your choice. You can use [https://imagemagick.org/index.php ImageMagick] to guess the Quality percentage, but
 +
keep in mind that it might be wrong, since the image might have been re-saved.
  
 
JPG PlusULTRA, while not considered successful in its initial purpose, may find some use on certain unique images.
 
JPG PlusULTRA, while not considered successful in its initial purpose, may find some use on certain unique images.
Line 1,302: Line 1,505:
 
DeJpeg Fatality PlusULTRA is its burly strongman big brother. Anything and everything that is JPG will tremble before you!
 
DeJpeg Fatality PlusULTRA is its burly strongman big brother. Anything and everything that is JPG will tremble before you!
  
Jaywreck by NMKD is trained to handle any amount of JPEG compression, even removing artifacts of quality 5 JPEGs. Its only weakness is a slight tendency to oversharpen certain textures. It's recommended to not use this model for very small images ( < 256px due to the large HR size it was trained on.
+
Jaywreck by NMKD is trained to handle any amount of JPEG compression, even removing artifacts of quality 5 JPEGs. Its only
 +
weakness is a slight tendency to oversharpen certain textures. It's recommended to not use this model for very small images
 +
( < 256px due to the large HR size it was trained on.
  
 
===DDS Files with BC1/DXT1, BC3/DXT5 Compression===
 
===DDS Files with BC1/DXT1, BC3/DXT5 Compression===
Line 1,334: Line 1,539:
 
|none (no interpolation)
 
|none (no interpolation)
 
|-
 
|-
|[https://drive.google.com/file/d/1LcBy7G3Xovgo38WySTQd1UM5PrINmJu1/view?usp=sharing BCGoneV2]
+
|[https://drive.google.com/file/d/1EYAoxC1tE2EhEgI4tiKujHUKjhq68gY-/view?usp=sharing BCGone_Smooth]
 
|[[User:Mutin Choler|Mutin Choler]]
 
|[[User:Mutin Choler|Mutin Choler]]
 
|1
 
|1
 
|[http://www.wtfpl.net/ WTFPL]
 
|[http://www.wtfpl.net/ WTFPL]
 
|BC1 Compression
 
|BC1 Compression
|225K
+
|110K
|1
+
|4
|128
+
|64
|250
+
|1400
|Custom (Photos)
+
|Images from DIV2K and random anime drawings
 
|[https://mega.nz/#!D940CAKR!-BAx7tSj1CgeopGkaMW31VuPsC3hvNBL35TEVAj4LIo 1xESRGAN]
 
|[https://mega.nz/#!D940CAKR!-BAx7tSj1CgeopGkaMW31VuPsC3hvNBL35TEVAj4LIo 1xESRGAN]
 
|-
 
|-
|[https://drive.google.com/file/d/1XS0uBDDhIpZGROFeASlfQrrTuUKc8k9r/view?usp=sharing BCGoneV3]
+
|[https://drive.google.com/file/d/1XS0uBDDhIpZGROFeASlfQrrTuUKc8k9r/view?usp=sharing BCGone_Detailed]
 
|[[User:Mutin Choler|Mutin Choler]]
 
|[[User:Mutin Choler|Mutin Choler]]
 
|1
 
|1
Line 1,359: Line 1,564:
 
|}
 
|}
  
BC1 (DXT1) compression is commonly used in dds textures, which are utilized in most PC games today, it allows to shrink the texture to 1/6 of the original size, reducing VRAM usage. There is also BC3 (DXT5) which uses BC1 compression for the color channels and leaves the alpha channel uncompressed, this one reduces the file size to 1/3 of the original. But this compression comes at a cost. The BC1 models are designed to reverse the damage done by the compression. This is a must if you want to upscale a dds file that uses either BC1 or BC3 compression. The free variant has more freedom when dealing with the images and should lead to better results. The restricted version only deals with perfect images compressed once (ideal case) so will perform worse in any other scenario, but it tries to preserve the original colors more. You can interpolate between restricted and unrestricted if you want.
+
BC1 (DXT1) compression is commonly used in dds textures, which are utilized in most PC games today, it allows to shrink
 +
the texture to 1/6 of the original size, reducing VRAM usage. There is also BC3 (DXT5) which uses BC1 compression for the
 +
color channels and leaves the alpha channel uncompressed, this one reduces the file size to 1/3 of the original. But this
 +
compression comes at a cost. The BC1 models are designed to reverse the damage done by the compression. This is a must if
 +
you want to upscale a dds file that uses either BC1 or BC3 compression. The free variant has more freedom when dealing with
 +
the images and should lead to better results. The restricted version only deals with perfect images compressed once (ideal
 +
case) so will perform worse in any other scenario, but it tries to preserve the original colors more. You can interpolate
 +
between restricted and unrestricted if you want.
  
 
===Cinepak, msvideo1 and Roq===
 
===Cinepak, msvideo1 and Roq===
Line 1,392: Line 1,604:
 
|}
 
|}
  
The Cinepak model removes movie compressions artifacts from older video compression methods like Cinepak, msvideo1 and Roq. Cinepak_alt may work better for video frames with realistic content.
+
The Cinepak model removes movie compressions artifacts from older video compression methods like Cinepak, msvideo1 and Roq.
 +
Cinepak_alt may work better for video frames with realistic content.
  
 
===H264 Compression===
 
===H264 Compression===
Line 1,499: Line 1,712:
 
|[https://mega.nz/#!vtgSWKQT!K7Asn2zKe4N70R2aV89KEMTKhH3aiyGAAiuQDJF09qs 2xESRGAN]
 
|[https://mega.nz/#!vtgSWKQT!K7Asn2zKe4N70R2aV89KEMTKhH3aiyGAAiuQDJF09qs 2xESRGAN]
 
|-
 
|-
|[https://drive.google.com/file/d/1Hxaa-uoccLInScJA9FOzfjAZBHtA90-0/view?usp=sharing DitherDeleter32]
+
|[https://drive.google.com/drive/u/1/folders/1XKREcpHnaYz_dcmkMPDbyETcXzDcQsR2 DitherDeleter]
 
|[[User:Mutin Choler|Mutin Choler]]
 
|[[User:Mutin Choler|Mutin Choler]]
 
|1
 
|1
 
|[http://www.wtfpl.net/ WTFPL]
 
|[http://www.wtfpl.net/ WTFPL]
|32 bit dedithering
+
|Dedithering
|90K
+
|104K
 
|4
 
|4
 
|64
 
|64
|800
+
|300
|Images from DIV2K
+
|Random anime images
|[https://mega.nz/#!D940CAKR!-BAx7tSj1CgeopGkaMW31VuPsC3hvNBL35TEVAj4LIo 1xESRGAN]
 
|-
 
|[https://drive.google.com/file/d/1B20zbC1s66q5PYAQqg9j-8aVJiR2rwOv/view?usp=sharing DitherDeleter64]
 
|[[User:Mutin Choler|Mutin Choler]]
 
|1
 
|[http://www.wtfpl.net/ WTFPL]
 
|64 bit dedithering
 
|70K
 
|4
 
|64
 
|800
 
|Images from DIV2K
 
 
|[https://mega.nz/#!D940CAKR!-BAx7tSj1CgeopGkaMW31VuPsC3hvNBL35TEVAj4LIo 1xESRGAN]
 
|[https://mega.nz/#!D940CAKR!-BAx7tSj1CgeopGkaMW31VuPsC3hvNBL35TEVAj4LIo 1xESRGAN]
 
|}
 
|}
  
Dithering is an older compression method, where the amount of colors gets reduced, if your image has few colors and have a noise like pattern in them try a De-Dither model.
+
Dithering is an older compression method, where the amount of colors gets reduced, if your image has few colors and havea noise
Ordered dithering is a less common form of dithering that results in distinctive checkerboard/crosshatch patterns, which are misinterpreted as texture by models not trained on it. It’s often used on GIFs because the pattern is stable between frames.
+
like pattern in them try a De-Dither model. Ordered dithering is a less common form of dithering that results in distinctive
 +
checkerboard / crosshatch patterns, which are misinterpreted as texture by models not trained on it. It’s often used on GIFs
 +
because the pattern is stable between frames.
  
4sFSDedither is made for photos/realistic images, but worth trying on other images that have reduced colors and dithering along with fine details. Trained using the ESRGAN-FS code for better details compared to plain ESRGAN.
+
4sFSDedither is made for photos/realistic images, but worth trying on other images that have reduced colors and dithering along
 +
with fine details. Trained using the ESRGAN-FS code for better details compared to plain ESRGAN.
  
4xFSDedither_Riven is a remake of 4xFSDedither that is fine-tuned to upscale images from the game Riven, but should be better in general, particularly on ordered dithering. I adjusted the dataset to have a better variety of dithering parameters, and turned up the HFEN and pixel loss to get better details and color restoration with less noise. 4xFSDedither_Riven_Smooth adds images with nearest-neighbor downscaling to the dataset for a smoother (but possibly less detailed) result.
+
4xFSDedither_Riven is a remake of 4xFSDedither that is fine-tuned to upscale images from the game Riven, but should be better in
 +
general, particularly on ordered dithering. I adjusted the dataset to have a better variety of dithering parameters, and turned
 +
up the HFEN and pixel loss to get better details and color restoration with less noise. 4xFSDedither_Riven_Smooth adds images with
 +
nearest-neighbor downscaling to the dataset for a smoother (but possibly less detailed) result.
  
 
====Banding====
 
====Banding====
Line 1,535: Line 1,742:
 
|-
 
|-
 
{{ModelDBColumns|Name|Author|Scale|License|Purpose|Iterations|Batch Size|HR Size|Dataset Size|Dataset|Pretrained Model}}
 
{{ModelDBColumns|Name|Author|Scale|License|Purpose|Iterations|Batch Size|HR Size|Dataset Size|Dataset|Pretrained Model}}
|-
 
|[https://drive.google.com/file/d/1wQ4QxTaVAszCnDuRPGOtbLoWcJHh0GIm/view DeBandizer]
 
|[[User:BlackScout|BlackScout]]
 
|1
 
|[https://www.gnu.org/licenses/gpl-3.0.en.html CC BY-NC 4.0]
 
|Images with Banding
 
|34k
 
|1
 
|128
 
|changing ~900
 
|Custom (Wallpapers)
 
|old attempts based on [https://mega.nz/#!D940CAKR!-BAx7tSj1CgeopGkaMW31VuPsC3hvNBL35TEVAj4LIo 1xESRGAN]
 
 
|-
 
|-
 
|[https://drive.google.com/file/d/1GNguVMV4E5t_4dn23w6UMXVKN6Ff9XPt/view?usp=sharing N64clean]
 
|[https://drive.google.com/file/d/1GNguVMV4E5t_4dn23w6UMXVKN6Ff9XPt/view?usp=sharing N64clean]
Line 1,559: Line 1,754:
 
|CC0 Textures
 
|CC0 Textures
 
|[https://mega.nz/#!D940CAKR!-BAx7tSj1CgeopGkaMW31VuPsC3hvNBL35TEVAj4LIo 1xESRGAN]
 
|[https://mega.nz/#!D940CAKR!-BAx7tSj1CgeopGkaMW31VuPsC3hvNBL35TEVAj4LIo 1xESRGAN]
 +
|-
 +
|[https://f002.backblazeb2.com/file/ESRGAN/_RELEASE/1x_Debandurh_FS_lite_140000_G.pth Debandurh_FS Ultra-lite]
 +
|[[User:Twittman|Twittman]]
 +
|1
 +
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 +
|Debanding images
 +
|140k
 +
|1/4
 +
|128
 +
|14.4k
 +
|Anime, real life, manga
 +
|None
 
|}
 
|}
  
Line 1,593: Line 1,800:
 
|}
 
|}
  
Just removes all colors and makes the image grayscale, the models here should restore some color to them, should also work for old black and white images.
+
Just removes all colors and makes the image grayscale, the models here should restore some color to them, should also work
 +
for old black and white images.
  
 
===Over-Sharpening===
 
===Over-Sharpening===
Line 1,613: Line 1,821:
 
|}
 
|}
  
The De-Sharpen model was made for rare particular cases when the image was destroyed by applying noise, i.e. game textures or any badly exported photos. If your image does not have any over-sharpening, it won’t hurt them, leaving as is. In theory, this model knows when to activate and when to skip, also can successfully remove artifacts if only some parts of the image are over-sharpened, for example in image consisting of several combined images, 1 of them with sharpen noise. It is made to remove sharpen noise, particularly made with Photoshop “sharpen” or “sharpen more” filters OR ImageMagick’s -sharpen directive with several varying parameters of Radius and Sigma, from subtle 0.3x0.5 to something extreme like 8x2, somewhere about that.
+
The De-Sharpen model was made for rare particular cases when the image was destroyed by applying noise, i.e. game textures
 +
or any badly exported photos. If your image does not have any over-sharpening, it won’t hurt them, leaving as is. In theory,
 +
this model knows when to activate and when to skip, also can successfully remove artifacts if only some parts of the image
 +
are over-sharpened, for example in image consisting of several combined images, 1 of them with sharpen noise. It is made to
 +
remove sharpen noise, particularly made with Photoshop “sharpen” or “sharpen more” filters OR ImageMagick’s -sharpen directive
 +
with several varying parameters of Radius and Sigma, from subtle 0.3x0.5 to something extreme like 8x2, somewhere about that.
  
 
===Aliasing===
 
===Aliasing===
Line 1,666: Line 1,879:
 
|Custom (?)
 
|Custom (?)
 
|none (no interpolation)
 
|none (no interpolation)
 +
|-
 +
|[https://u.pcloud.link/publink/show?code=kZEdNHXZ3bpVwUoge94d9igFWxbjQQrcbpny NormalMapGenerator-CX Lite]
 +
|Joey
 +
|1
 +
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 +
|Generating normal maps from textures
 +
|200k
 +
|Varied
 +
|64
 +
|14.784K
 +
|CC0 textures/normals, randomly downscaled OTF (by 1, 2, and 4) to get more variation
 +
|DIV2K-Lite
 +
|-
 +
|[https://u.pcloud.link/publink/show?code=kZecVzXZEO6fAyO6w25YDc43BkSVHRgFdOwX FrankenMapGenerator-CX Lite]
 +
|Joey
 +
|1
 +
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 +
|Generating roughness and displacement maps
 +
|215k
 +
|Varied
 +
|64
 +
|14.784K
 +
|CC0 textures/normals, randomly downscaled OTF (by 1, 2, and 4) to get more variation
 +
|DIV2K-Lite
 
|}
 
|}
  
Line 1,790: Line 2,027:
 
https://drive.google.com/drive/folders/1ldwajXL50uC7PCS63B4Wato6Dnk-svNL
 
https://drive.google.com/drive/folders/1ldwajXL50uC7PCS63B4Wato6Dnk-svNL
  
These models were transformed from the original RRDB_ESRGAN_x4.pth model into the other scales, in order to be used as pretrained models for new models in those scales.
+
These models were transformed from the original RRDB_ESRGAN_x4.pth model into the other scales,
 +
in order to be used as pretrained models for new models in those scales.
  
 
More information can be found [https://github.com/victorca25/BasicSR/wiki/Interpolation here].
 
More information can be found [https://github.com/victorca25/BasicSR/wiki/Interpolation here].
Line 1,806: Line 2,044:
 
|?
 
|?
 
|?
 
|?
 +
|none
 +
|-
 +
|}
 +
 +
=SOFVSR ("vicGAN") Models=
 +
 +
==Real World==
 +
 +
{| class="wikitable sortable"
 +
|-
 +
!Name!!Author!!Scale!!License!!Purpose!!Iterations!!Batch Size!!HR Size!!Dataset Size!!Dataset!!Architecture!!Num Frames!!Pretrained Model
 +
|-
 +
|[https://u.pcloud.link/publink/show?code=kZojwLXZfeqA3Eoyv6SKXXt01AlSeuHH0WKX 4x_REDSVAL-7f-RRDB-Lite]
 +
|Joey
 +
|4
 +
|CC BY-NC-SA 4.0
 +
|Upscaling real-world video / Pretrained model for 7-frames RRDB
 +
|200k
 +
|Varied
 +
|128
 +
|3,000
 +
|REDS validation set
 +
|RRDB
 +
|7
 +
|none
 +
|-
 +
|}
 +
 +
==Cartoons==
 +
 +
 +
{| class="wikitable sortable"
 +
|-
 +
!Name!!Author!!Scale!!License!!Purpose!!Iterations!!Batch Size!!HR Size!!Dataset Size!!Dataset!!Architecture!!Num Frames!!Pretrained Model
 +
|-
 +
|[https://u.pcloud.link/publink/show?code=XZMmozXZjW4hHqJu7ikbQpyeqUItTh7QJO6k 3x_Video_TSSM]
 +
|Joey
 +
|3
 +
|CC BY-NC-SA 4.0
 +
|Upscaling cartoon DVDs
 +
|70k
 +
|16
 +
|192
 +
|125,436
 +
|The Spongebob Squarepants Movie
 +
|SOFVSR net
 +
|3
 +
|none
 +
|-
 +
|[https://u.pcloud.link/publink/show?code=XZ2AERXZ9YpzzV4rRvySFiq3MsrTEYdEjtkV 3x_Video_TSSM_RRDB]
 +
|Joey
 +
|3
 +
|CC BY-NC-SA 4.0
 +
|Upscaling cartoon DVDs
 +
|80k
 +
|6
 +
|192
 +
|125,436
 +
|The Spongebob Squarepants Movie
 +
|RRDB
 +
|3
 +
|none
 +
|-
 +
|[https://u.pcloud.link/publink/show?code=kZ5QKLXZ5mPYmsGpGU75DLNII9AOlYQNqMjk 2x_SBS11-RRDB]
 +
|Joey
 +
|2
 +
|CC BY-NC-SA 4.0
 +
|Upscaling cartoon DVDs with bad haloing and noise
 +
|111k
 +
|6
 +
|128
 +
|126,956
 +
|Every 6 frames from every episode of Spongebob Season 11 (DVD and HD). DVD was resized with Spline36 to 540p. LR also got extra OTF JPEG (1/4 chance) and sharpening (15% chance)
 +
|RRDB
 +
|3
 +
|none
 +
|-
 +
|}
 +
 +
==Pretrained models==
 +
 +
{| class="wikitable sortable"
 +
|-
 +
!Name!!Author!!Scale!!License!!Purpose!!Iterations!!Batch Size!!HR Size!!Dataset Size!!Dataset!!Architecture!!Num Frames!!Pretrained Model
 +
|-
 +
|[https://mega.nz/file/28JmyLrK#xhRP-EZKR7Vg7UjIRZQqotiFLix21JaGGLSvZq7cjt4 video_G]
 +
|Victorca25
 +
|4
 +
|?
 +
|Pretrained model for regular SOFVSR models
 +
|?
 +
|32
 +
|32
 +
|REDS' size
 +
|REDS
 +
|SOFVSR net
 +
|3
 +
|none
 +
|-
 +
|[https://mega.nz/file/TwwEWD7Q#wCfUvVudI17weYc1JLeM3nTeK2xiMlVdc_JN1Nov3ac VESRGAN_G]
 +
|Victorca25
 +
|4
 +
|?
 +
|Pretrained model for RRDB SOFVSR models
 +
|?
 +
|32
 +
|32
 +
|REDS' size
 +
|REDS
 +
|RRDB
 +
|3
 
|none
 
|none
 
|-
 
|-
 
|}
 
|}

Revision as of 18:13, 10 February 2021

If a model has no license, that means all rights reserved. You will need a private license or explicit permission from the creator to use it commercially or with modifications. However you may use all model here on non-commercial projects, that may not in any way cost money. If you gain permission to use a model in another way including commercially, you should still credit the author and link to the original model source.

ESRGAN ("old Architecture") Models

Models that use the "old" ESRGAN architecture. They can be used either with the official ESRGAN repository (old arch tag) and BasicSR (old arch or preferably victorc's fork). You can also use tools that wrap around one of them, like IEU from Honh.

Image scaling and Video upscaling

Drawings

Manga / Anime

Name
Author
Scale
License
Purpose
Iterations
Batch Size
HR Size
Dataset Size
Dataset
Pretrained Model
"
Deviance V1 BlackScout 4 GNU GPLv3 Digital Art Upscaler 60K 12 128 5.435K Digital Art from DeviantArt 4xESRGAN
Deviance V2 BlackScout 4 GNU GPLv3 Digital Art Upscaler 82K 8 128 0.676K Custom MipMapped 4xESRGAN
FireAlpha BlueAmulet 4 CC BY-NC 4.0 Artwork with an alpha channel 1.285M 128 104k tiles, 1.2k imgs Fire Emblem artwork with transparency None
WaifuGAN v3 DinJerr 4 CC BY-NC 4.0 Anime / Manga 30K 2 128 0.173K CG-Painted Anime Manga109v2
Dedither Manga FS Jacob_ (buildist) 4 GNU GPLv3 Manga and pixel art 200K 16 128 1,526 Manga109 copied 14 times with different dithering and scaling 4xESRGAN
Manga109Attempt Kingdomakrillic 4 CC BY 4.0 Anime / Manga ? 4 ? 0.1K Manga109 4xPSNR
De-Toon LyonHrt 4 CC BY-NC-SA 4.0 Toon Shading / Sprite 225K 8 128 7.117K Custom Cartoon-style photos 4xPSNR
Falcon Fanart LyonHrt 4 CC BY-NC-SA 4.0 Anime / Manga 125K 8 128 3.393K Falcon Fanart 4xPSNR
Fidelbd Pokemodel Neo-Raws#4055 2 WTFPL Old Anime 300K 4 128 10.242K Pokemon DVD - BluRay 2xESRGAN
Pokemodel Lite Neo-Raws#4055 2 WTFPL Old Anime 100K 4 128 268 Pokemon 2xESRGAN
DeviantPixelHD Raulsangonzalo 4 CC BY-NC 4.0 Digital Art Upscaler 250K 16 128 2.797K Digital Art from Deviant Art 4xPSNR
Anime Oldies solidd93110#7731 4 WTFPL upscaling old anime 290K 16 128 31.316K Frames from Anime BluRay 4xPSNR
Anime Oldies Alternative solidd93110#7731 4 WTFPL captain tsubasa anime 290K 16 128 1.250K Frames from Anime BluRay 4xPSNR
Sharp Anime V1 solidd93110#7731 2 CC BY-NC-SA 4.0 Lines and Details 300K 4 128 7.738K Anime DVD - BluRay 2xPSNR
DigiPaint TheAtheistGod 4 CC0 Digital Art Upscaler 35K 6 128 48.493K Misc digital art, material studies 4xfalcoon300(manga)
2x_Byousoku_5_Centimeter Mystery_Bullet#0642 (Shady Adel) 2 GNU GPLv3 Anime landscape upscale 300k 8 128 3K Frames from BluRay of Byousoku 5 Centimeter 2xESRGAN
UltraYandere nmkd 4 WTFPL Highly flexible 2D Art upscaling 300k 3 256 4800 unique, 28800 total (augmented) Yandere4800 (yande.re artworks) 4xESRGAN
PatchySharp nmkd 4 WTFPL Upscaler for clean images or images with compression artifacts 240k 5 256 1800 DIV2K Clean + JPEG 4xESRGAN

Manga109Attempt is slightly blurry, but performs well as a general upscaler.

Falcoon Fanart tries to improve upon it with the goal of removing checkerboard patterns / and dithering. It has oil color based shading with sharp lines.

WaifuGAN v3 is DinJerr’s third attempt at training from a mostly anime dataset sourced from image boards and is intended for upscaling CG-painted anime with variable outlines. Only PNGs were used, mainly with brush strokes and gradients. Texturised images avoided as much as possible. If too generative, tone down by interpolating with a softer model.

De-Toon, is a model that does the opposite of tooning an image. It takes toon style shading and detail, and attempts to make it realistic. Its very sensitive, and can be used on small sprites, to large images. Also included is a alt version, which is less sharp.

DigiPaint is a digital art upscaler designed to take brush strokes into account, based off of the Falcoon Fanart model.

DeviantPixelHD was designed to upscale original LucasArts games' backgrounds. High Definition digital art from Deviant Art was used, reducing the LR training to 32 colors so the training images looked pixelated. Can be used with other environments such as manga or digital art. Examples can be found on Raúl Sangonzalo's YouTube account (https://www.youtube.com/channel/UCwBfuiHdSPQ-zslOmiW8OHg).

FireAlpha is an attempt to create a 4 channel model that handles transparency without splitting and upscaling separately.

  • Notice: IEU can use the model but will attempt to split and handle transparency itself, BlueAmulet's fork of ESRGAN is required to use this model. https://github.com/BlueAmulet/ESRGAN

4xFSDedither_Manga was intended for pixel art containing dithering patterns, but ended up being effective on pixel art and manga/cartoon style images without any dithering at all.

Cartoon / Comic

Name
Author
Scale
License
Purpose
Iterations
Batch Size
HR Size
Dataset Size
Dataset
Pretrained Model
"
Comic Book LyonHrt 4 CC BY-NC-SA 4.0 Comic / Drawings 115K 8 128 1.548K Custom (Spider-Man) none (no interpolation)
DeteriorationFrames 1.0 Klexos 1 CC BY-NC-SA 4.0 Animation 102K 2 128 6.75K HD to bad DVD transfer Scratch
FilmFrames 0.5 Klexos 2 CC BY-NC-SA 4.0 Film Animation 677K 3 192 250 SD to HD film rescan Scratch
CelFrames 1.0 Klexos 4 CC BY-NC-SA 4.0 Film Animation 20K 38 128 30k 4K Animation Cel Scans DigitalFrames 2.0
DigitalFrames 2.0 Klexos 4 CC BY-NC-SA 4.0 Digital Animation 905K 27 128 1.636M Digital Cartoon Images Scratch
DigitalFrames 2.1_Aggressive Klexos 4 CC BY-NC-SA 4.0 Flat Digital Animation 230K 27 128 1.8M Digital Cartoon Images DigitalFrames 2.0
DigitalFrames 2.5 Klexos 4 CC BY-NC-SA 4.0 Digital Animation 925K 38 128 30K SD to HD Cartoons DigitalFrames 2.0
Struzan Laserschwert 4 CC BY-NC-SA 4.0 airbrush/pencil-based artwork 300K 4 128 2.790K movie poster artwork by Drew Struzan and Bill Eaken 4xPSNR
FatalimiX Twittman 4 CC BY-NC-SA 4.0 Comic and Cartoon style images 260K 10 128 69k/79k Digital Comics 4x_Fatality_MKII_90000_G_02.pth
SpongeBob Joey 4 CC BY-NC-SA 4.0 Cartoons & SpongeBob Games 235k 20 128 8k Scene-detected frames from random Season 11 episodes 4xESRGAN
SpongeBob v6 Joey 4 CC BY-NC-SA 4.0 Cartoons 190k 20 128 4,803 1 frame of every scene of season 11, downscaled 50% with nearest neighbor 4xESRGAN
Spongebob v6 De-Quantize Joey 4 CC BY-NC-SA 4.0 Quantized Cartoons 90k 20 128 4,803 1 frame of every scene of season 11, downscaled 50% with nearest neighbor 4x Spongebob v6
Spongebob v6 Deblur Joey 4 CC BY-NC-SA 4.0 Cartoons 65k 20 128 4,803 1 frame of every scene of season 11, downscaled 50% with nearest neighbor 4x Spongebob v6
HugePeeps DinJerr 8 CC BY-NC 4.0 Painted humans 360k 4 256 1,000+ A bunch of pretty girls (and guys) from ArtStation TGHQFace8x
Fatal_Anime Twittman 4 CC BY-NC-SA 4.0 Anime and Cartoons 500k 6 128 183k Anime and Manga images 4x_Fatality_Comix_MKII_430000_G.pth
HugePaint DinJerr 8 CC BY-NC 4.0 Digital Illustrations 500k 4 256 1,000+ Variety of images from ArtStation HugePeeps
4x_AbeScale mitch#1337 4 GNU Linework Cartoons 110k 8 192 8 LR: Clone High DVD, HR: Re-illustrated Vectors from Frames 4xESRGAN

The Comic Book model was trained using stills from the film spiderman into the spiderverse, has a comic book crosshatch shading effect to the images. Sample

FilmFrames 0.5: Purpose of the model is upscaling medium quality sourced SD animation frames to an HD "Film rescan" status. https://imgsli.com/MjA2NDM

CelFrames 1.0: Trained on thousands of 4K high quality cel scans, the purpose of this model is restore the frame to it's original Cel look, with a particular emphasys on textures and preserving those brushstrokes and natural lines. Not to be used with heavily grainy sources. https://imgsli.com/Mzk2NzU

DeteriorationFrames 1.0: The purpose of this model is to aid in the creation of datasets aimed at animation by saving you time and allowing you to deteriorate the HQ/HD frames directly though this model instead of having to go through frame by frame to find SD/LQ counterparts. This iteration in particular adds dot crawl, ringing and chromatic aberrations. https://imgsli.com/MTk4MjA. Important note: Resizing your HD counterparts before using this model to create the LRs is essential. Tip: If you blur the HRs after resizing them but before using the deterioration model on them, depending on how much you blur you'll get less Chromatic aberrations. So with this you can tweak both blurriness and the aberrations to match better the source you intend to actually use the model you are about to train on.

DigitalFrames 2.0: Was trained completely separately from 1.0 with a completely new and bigger dataset. It has a reduced sensitivity to noise which reduces the blobs of noise phenomena. It can also work reasonably well on Cel/Film Toons now. Overall it's a big leap from 1.0.

DigitalFrames 2.1_Aggressive: Provides the sharpest results yet. It's the most aggressive version of DF. Does it's best on animation with little to no textures. The model as well as any version of DF is not suitable for all kinds of animation.

DigitalFrames 2.5: Big leap from 2.0, especially as much as texture retention is concerned https://imgsli.com/Mzk2MzU.

FatalimiX is specialized in making those low res comics high res again!

The SpongeBob model was trained for the purpose of upscaling Battle for Bikini Bottom's textures, but it works well for upscaling scenes from the show as well. It was initially trained on the clean downscaled images but later JPEG, quantization, dithering, and Gaussian blur were added to the LR OTF options to help it upscale the textures better.

Spongebob v6 is ideally a lot sharper and cleaner, but Joey's still not sure if it works better than the old one. It still seems to be better in many cases.

Spongebob v6 De-Quantize is for de-quantizing as well as upscaling. The results are pretty blurry but it works decently enough.

Spongebob v6 Deblur was created as Joey wasn't entirely happy with the main Spongebob v6 model. It was trained with blurring OTF options and two different downscale types.

HugePeeps v1 was created by DinJerr to test out a theory on how to properly train 8x models. This particular model best works on painted pics of ladies.

HugePaint v1 is trained on a more generic dataset with a balance of illustrated humans, vegetation, minerals and some animals.

Abescale is an upscale model intended to bring Clone High DVDs (480i/p) to 4k. Based on re-illustrated frames from the DVD, is effective on similar cartoons with heavy linework.

Some of the first cartoon painted models for ESRGAN can still be found here, but have been mostly superseded by newer models.

Pixel Art

used in tandem with other bitmaps to construct a larger scene.

Name
Author
Scale
License
Purpose
Iterations
Batch Size
HR Size
Dataset Size
Dataset
Pretrained Model
"
BigFace v3 DinJerr 4 CC BY-NC 4.0 Painted people 300K~ 4 384 1K Selected concept art BigFace
BigFArt Suite DinJerr 4 CC BY-NC 4.0 Larger-scaled pixels to digital painting 800+ 6 192 ~12K DIV2K, FatalityFaces, ArtStation paintings Face Focus
FArtDIV3 Suite DinJerr 4 CC BY-NC 4.0 Pixels to digital painting 700K~ 4 128 ~10K DIV2K, FatalityFaces, ArtStation paintings ArtStation1337
Faithful Joey 2 CC0 General small pixel textures 130K 80 32 2.858K Faithful 32x32 (HR) / default Minecraft 1.13 and 1.14 (LR) Private model trained from 2xESRGAN
Faithful Lite Joey 2 CC BY-NC-SA 4.0 General small pixel textures 275K 128 32 2.928K Faithful 32x32 1.13-1.16(HR) / default Minecraft 1.13 and 1.16 (LR) none
FakeFaith Lite Joey 2 CC BY-NC-SA 4.0 General small pixel textures 105K 24 64 24.110K Spongebob Season 11 frames, downscaled 50% with NN for the HR, and 50% NN again for the LR Faithful Lite
SmoothRealism Joey 4 CC BY-NC 4.0 Pixel art, rocky/grainy textures 140K 2 64 679 R3DCraft Smooth Realism Texture Pack 64x 4xPSNR
Rebout LyonHrt 4 CC BY-NC-SA 4.0 Character Sprites 325K 8 128 23.808K Custom prepared sprites from kof 94 rebout, now updated with cx loss Detoon
Rebout blend LyonHrt 4 CC BY-NC-SA 4.0 Character Sprites 325K 8 128 23.808K Custom prepared sprites from kof 94 rebout, which with gradients blending Detoon
ScaleNX LyonHrt 4 CC BY-NC-SA 4.0 Scalenx style pixel art upscaler 80K 8 128 1.070K custom scalenx up-scaled from retroarch shader 4xPSNR
Xbrz LyonHrt 4 CC BY-NC-SA 4.0 Xbrz style pixel art upscaler 90K 8 128 1.897K custom xbrz up-scaled 4xPSNR
Xbrz+DD LyonHrt 4 CC BY-NC-SA 4.0 Xbrz style pixel art upscaler with de-dithering 90K 8 128 1.523K custom de-dithered xbrz xbrz
Fatality Twittman 4 CC BY-NC-SA 4.0 (dithered) sprites 265K 10 128 19.7K ? link
Arzenal (v1.1) Glasshopper 8 CC BY-NC-SA 4.0 Smooth general pixel art, Minecraft textures 175K ? ? ? None used, model was interpolated nmkd's pixel art upscaling models with different interpolation settings and de-dithered versions
Fatality MK2 Twittman 4 CC BY-NC-SA 4.0 Sprites 90K 10 128 42K Anime, Manga and some real life A previously attempted MK2
Fatal Pixels Twittman 4 CC BY-NC-SA 4.0 Sprites 340K 22 96 49K Anime, Manga Fatality_MKII_90k
PixelPerfect Mutin Choler 4 WTFPL Sprites 137K 4 128 300 Anime, Manga 1xESRGAN

Fatality is meant to be used for upscaling medium resolution Sprites, dithered or un-dithered, it can also upscale manga/anime and gameboy camera images.

Arzenal is a mashup of multiple of nmkd's pixel art models, including de-dithered version for noise reduction, specifically designed to work on many different kinds of pixel art, without being too smooth, or too rough. The reason being is that nmkd's models are good, but they are TOO good for their purposes, and a model that evens them out can create results that fit both tasks. The de-dithering was also added for smoothing, and noise reduction.

Rebout is trained to give detail to character models, with faces and hands improved. Based on the snk game kof94 rebout, although best for snk style games, does work on a variety of sprites. Also included is a interpolated version that may provide a cleaner upscale for certain sprites.

Faithful 2x was trained on the faithful 32x32 Minecraft texture/resource pack, using both the 1.13 and 1.14 versions (which have different textures). For those not familiar, this texture pack is meant to upscale the 16x16 textures to 32x32 faithfully, as in make them as close as possible to the original, but higher resolution. The model does a decent job at replicating this.

  • NOTE: Due to an error in the dataset, sometimes the model turns jagged white lines green. Only encountered this in one texture

tried, so it appears unlikely to happen.

FArtDIV3 Suite is a group of models based on a large image dataset trained using OTF tile generation. Most of the models are based off the Base model, and suite different purposes. Fine is used to generate details, but may be noisy/grainy at times. Blend will attempt to blend colour steps into proper gradients. UnholyFArt is a interpolated model that reduces details in favor of sharp blends.

BigFArt Suite is similar to FArtDIV3, but for larger pixel sprites/art in the context of tile size.

BigFace_v3 Suite takes the BigFace model from the BigFArt suite, and tries to improve by using double-sized tiles. Main model has good dedithering. _Blend focuses on sharp blends, and _Clear will abstract and sharpen details.

Photographs and Photorealism

Misc / Kitchen Sink

All kinds of photographs or photorealistic images. Those models aren't specialized.

Name
Author
Scale
License
Purpose
Iterations
Batch Size
HR Size
Dataset Size
Dataset
Pretrained Model
"
Misc Alsa 4 CC BY-NC-SA 4.0 Surface Textures 220K 32 128 20.797K Custom (Photos) Manga109Attempt
Nickelback BlackScout 4 GNU GPLv3 Photographs 70k 8 128 44.55k Wallpapers 4xESRGAN
Fuzzy Box BlueAmulet 4 CC BY-NC 4.0 Anything 197K 16 128 9.962K CC0 Textures 4xESRGAN
Box Jacob_ (buildist) 4 GNU GPLv3 Realistic 390K 8 192 11.577K Flickr2K+Div2K+OST PSNR model from same data
Lollypop LyonHrt 4 CC BY-NC-SA 4.0 Universal 235k 8 128 109.5K prerendered images 4xESRGAN
UniversalUpscaler-Soft Mutin Choler 4 WTFPL Universal 190K 4 128 1400K DIV2K + Anime 4xESRGAN
UniversalUpscaler-Detailed Mutin Choler 4 WTFPL Universal 155K 4 128 1400K DIV2K + Anime 4xESRGAN
Superscale nmkd 4 WTFPL Clean Real-World Images 134k 1/2 448 900 DIV2K Train + Valid 4xESRGAN
Ground ZaphodBreeblebox 4 Ground Textures 305K ? 128 ? Custom (Ground textures Google) ?

Box was meant to be an improvement on the RRDB_ESRGAN_x4 model (comparison). It’s also trained on photos, but with a much larger dataset which was downscaled with linear interpolation (box filter) instead of bicubic.

Deoldify is intended for denoising and deblurring photos. All outputs are greyscale. This was initially a proof of concept. (Need clarification on specific subject matter.)

The Ground model was trained on various pictures of stones, dirt and grass using Google’s image search.

The Misc model is trained on various pictures shoot by myself, including bricks, stone, dirt, grass, plants, wood, bark, metal and a few others.

Superscale by NMKD was made to rival the RealSR models and is one of the best upscalers for real-world content. It's trained on PNGs, so it doesn't handle compression artifacts well.

Lollypop was originally made for prerendered manga faces, using the patchgan discriminator. Combined with cutmixup, cx loss, and Frequency Separator, it handles prerendered, realistic faces, and even pixel art, with dedithering low palette images incredibly well.

Characters and Faces

For images of humans, creatures, faces, ...

Name
Author
Scale
License
Purpose
Iterations
Batch Size
HR Size
Dataset Size
Dataset
Pretrained Model
"
Trixie LyonHrt 4 CC BY-NC-SA 4.0 Star Wars 275K 8 192 19.814K ? None
Face Focus LyonHrt 4 CC BY-NC-SA 4.0 Face De-blur 275K 8 192 4.157K Custom (Faces) 4xPSNR
Face Twittman 4 CC BY-NC-SA 4.0 Face Upscaling 250K 10 128 3.765K Custom (Faces) 4xESRGAN
Face-Ality V1 Twittman 4 CC BY-NC-SA 4.0 Face Upscaling 310k 10 128 13.3k Custom (Faces) 4x_Faces_04_N_180000_G.pth
Cat_Patch Twittman 4 CC BY-NC-SA 4.0 For upscaling Meow 325k 8 128 9.5k Custom (Cats) previous attempt
TGHQFace8x Torrentguy 8 GNU GPLv3 Face Upscaling 500k 8 128 70k Flickr Cropped Faces 8xESRGAN

Trixie was made to bring balance to the force… Also to upscale character textures for star wars games, including the heroes, rebels, sith and imperial. Plus a few main aliens…Why called trixie? Because jar jars big adventure would be too long of a name… This also provides good upscale for face textures for general purpose as well as basic star wars textures.

The Face Focus modes was designed for slightly out of focus / blurred images of faces. It is aimed at faces / hair, but it can help to improve other out of focused images too as always just try it.

Specialized

For Purposes that didn't fit anywhere else (for now).

Name
Author
Scale
License
Purpose
Iterations
Batch Size
HR Size
Dataset Size
Dataset
Pretrained Model
"
Skyrim Armory Alsa 4 CC BY-NC-SA 4.0 Armor, Clothes and Weapons 80K 26 128 0.8K Skyrim Mod textures Manga109Attempt
Nickelfront BlackScout 4 GNU GPLv3 Money, Coins 14k 8 128 1.176k Images of Coins 4xESRGAN
BS_Screenbooster_SPSR BlackScout 4 GNU GPLv3 Upscale game screenshots 76K 2 128 1.2M (tiles) Game screenshots 4xESRGAN
Wolly BlackScout 2 GNU GPLv3 Pixar Movies / Wall-E 38k 1 256 9.0k BD 1080p Wall-E Frames 2xESRGAN
Fallout 4 Weapons Bob 4 CC BY-NC-SA 4.0 Fallout Weapon Diffuse Textures 120K 13 128 532(OTF) Fallout 4 HDDLC Weapon Diffuse Textures Manga109Attempt
Fallout Weapons V2 Bob 4 CC BY-NC-SA 4.0 Video game textures, mostly metal rusty, clean or painted. 180K 13 128 1.999K(OTF) Fallout 4 HD DLC(Weapon,Armor/Clothes,Vehicle and Interior/Architecture textures to be exact) Fallout 4 Weapons
Skyrim Misc Deorder 4 CC0 Skyrim Diffuse Textures 105K ? 128 ? Skyrim Diffuse Textures ?
PackCraft V4 Joey 4 CC BY-NC-SA 4.0 Upscaling Minecraft pack.png 40K 1 512 60K Alpha Minecraft 4xESRGAN
Skyrim Wood Laeris 4 Wood 75K ? ? ? ? ?
Forest LyonHrt 4 CC BY-NC-SA 4.0 Wood / Leaves 160K 8 192 2.2K Custom(?) none
Map LyonHrt 4 CC BY-NC-SA 4.0 Map / Old Paper with text 120K 8 192 2.311K Custom(Scans) none
Morrowind Mixed mkultra 4 CC BY-NC-SA 4.0 Morrowind Mod Textures 115K-130K 16 128 56K Morrowind Mod Textures 4xESRGAN

The map model was trained on maps, old documents, papers and various styles of typefaces/fonts. Based on a dataset contributed by alsa. Sample

The Forest model is focused on trees, leaves, bark and stone can be used for double upscaling for even more detail. Sample

The Armory model was trained with modded textures form Skyrim, including Clothing, Armor and Weapons. (Leather, Canvas and Metal should all work - maybe too sharp so interpolate)

The wood model was trained for Skyrim by Laeris.

The Skyrim Diffuse models is supposed to be used with Skyrim’s diffuse textures. It is a bit too sharp so I recommend to interpolating with the |4xESRGAN model or the mangaAttempt109 model, look in Deorder’s Skyrim Model Google Drive for an already interpolated version.

Fallout 4 weapons was trained using Fallout 4’s official HDarmor/weapon textures but could be used on other weapon and armor textures.

Grayscale

Name
Author
Scale
License
Purpose
Iterations
Batch Size
HR Size
Dataset Size
Dataset
Pretrained Model
"
Skyrim Alpha Deorder 4 CC0 Alpha Channel 105K ? 128 ? Alpha Channels from Skyrim ?
Deoldify TheAtheistGod 4 CC0 B/W Photo Restoration 790K 16 128 2.931M ADE20K, DIV2K, Library of Congress Falcon Fanart (?)

Trained to upscale grayscale images, like specular or alpha etc.

Artifact Removal

The models in this section were made to remove compression artifacts in images and textures.

Blur

Name
Author
Scale
License
Purpose
Iterations
Batch Size
HR Size
Dataset Size
Dataset
Pretrained Model
"
LADDIER1 Alexander Syring 4 CC0 Remove noise, grain, box/lens/guassian blur 282K 8 128 12.7K Mostly images of Nirvana, some of other images 4xESRGAN
Fatality DeBlur Twittman 1 CC BY-NC-SA 4.0 Deblurring 270K 1 128 41K Mix of anime, manga, and photos 1x_DeJpeg_Fatality_01_175000_G.pth

Fatality DeBlur is intended for general deblurring of images, and has some robustness against compression and noise.

LADDIER1 upscales images with noise, grain and different types of blur.

Noise

Name
Author
Scale
License
Purpose
Iterations
Batch Size
HR Size
Dataset Size
Dataset
Pretrained Model
"
ISO denoise v1 Alpha 1 CC0 Remove high ISO noise 372K 1 128 170K Custom (DSLR photos) 1xESRGAN
ISO denoise v2 Alpha 1 CC0 Remove high ISO noise 308K 1 128 239K Custom (DSLR photos) ISO denoise v1
NoiseToner (Uniform) Mutin Choler 1 WTFPL Remove Uniform noise 100K 4 64 1.4K DIV2K + Anime 1xESRGAN
NoiseToner (Poisson) Mutin Choler 1 WTFPL Remove Poisson noise 150K 4 64 1.4K DIV2K + Anime 1xESRGAN

High ISO is required to capture images at very high shutter speeds and/or in poor lighting. However, increasing ISO also results in noise that is challenging to remove without also losing fine details. To remove this noise without damaging details, one can apply the technique of median stacking: combine 10-100 photos of the exact same object to separate the signal (clear photo) from noise.

ISO denoise v1 is trained on a handcraft set of pairs between original noisy photos and median stacked counterparts. It tends to be conservative, avoiding modifying anything that does not closely resemble camera ISO noise. ISO denoise v2 uses a larger dataset, handling some cases better and some cases worse than v1. If results are unsatisfactory with one model, try the other model.

JPG Compression

Name
Author
Scale
License
Purpose
Iterations
Batch Size
HR Size
Dataset Size
Dataset
Pretrained Model
"
jpg (0-20%) alsa Alsa 1 CC BY-NC-SA 4.0 JPG compressed Images 178K 2 128 6.23K Custom (Photos / Manga) JPG (20-40%)
jpg (20-40%) alsa Alsa 1 CC BY-NC-SA 4.0 JPG compressed Images 141K 2 128 6.23K Custom (Photos / Manga) JPG (40-60%)
jpg (40-60%) alsa Alsa 1 CC BY-NC-SA 4.0 JPG compressed Images 100K 2 128 ~6.5K Custom (Photos / Manga) JPG (60-80%)
jpg (60-80%) alsa Alsa 1 CC BY-NC-SA 4.0 JPG compressed Images 91K 2 128 ~6.5K Custom (Photos / Manga) JPG (80-100%)
jpg (80-100%) alsa Alsa 1 CC BY-NC-SA 4.0 JPG compressed Images 162K 2 128 ~6.5K Custom (Photos / Manga) BC1 take 1
jpg (20-40%) BlueAmulet BlueAmulet 1 CC BY-NC 4.0 JPG compressed Images 535K 1 128 52.789K Custom (CC0 Textures) JPG (40-60%)
jpg (0-20%) BlueAmulet BlueAmulet 1 CC BY-NC 4.0 JPG compressed Images 610K 1 128 52.789K Custom (CC0 Textures) JPG (20-40%)
jpg (40-60%) BlueAmulet BlueAmulet 1 CC BY-NC 4.0 JPG compressed Images 550K 1 128 52.789K Custom (CC0 Textures) JPG (60-80%)
jpg (60-80%) BlueAmulet BlueAmulet 1 CC BY-NC 4.0 JPG compressed Images 545K 1 128 52.789K Custom (CC0 Textures) JPG (80-100%)
jpg (80-100%) BlueAmulet BlueAmulet 1 CC BY-NC 4.0 JPG compressed Images 550K 1 128 52.789K Custom (CC0 Textures) BC1 take 1
Jaypeg90 Jacob_ 4 GNU GPLv3 JPG Photos 85-95% Quality 66K 24 128 3.440K * 2 Flickr2K+Div2K, 85-95% JPEG Compressed, cubic or NN scaling 4xESRGAN
SBDV-DeJPEG-Lite Joey 1 CC BY-NC-SA 4.0 JPG compressed Images 130k 12 64 934 SpongeBob, DIV2K 1x_DIV2K-Lite_80k.pth
Jaywreck nmkd 1 WTFPL JPG compressed Images 190k 1/2 132 900 DIV2K Train + Test 1xESRGAN
Jaywreck3 nmkd 1 WTFPL JPG compressed Images 320k 4 112 1800 (900 unique) DIV2K Train + Valid 1xESRGAN
JPG PlusULTRA Twittman 1 CC BY-NC-SA 4.0 JPG compressed Images 130K 1 ? 0.937K Custom (Manga) Failed Attempts
DeJpeg Fatality PlusULTRA Twittman 1 CC BY-NC-SA 4.0 JPG compressed Images 200K 1 128 26K Real life, Manga and digital text 1x_DeJpeg_Fatality_01_200000_G.pth
SaiyaJin DeJpeg Twittman 1 CC BY-NC-SA 4.0 JPG compressed Images from 2% to 100% quality 300K 1/8 128 22.8K Anime, real life, manga, generated 1x_DeJpeg_Fatality_PlusULTRA_200000_G.pth

JPG gets compressed with a Quality Percentage between 0 and 100. So depending on how bad your JPEGs are compressed, choose the model of your choice. You can use ImageMagick to guess the Quality percentage, but keep in mind that it might be wrong, since the image might have been re-saved.

JPG PlusULTRA, while not considered successful in its initial purpose, may find some use on certain unique images.

DeJpeg Fatality PlusULTRA is its burly strongman big brother. Anything and everything that is JPG will tremble before you!

Jaywreck by NMKD is trained to handle any amount of JPEG compression, even removing artifacts of quality 5 JPEGs. Its only weakness is a slight tendency to oversharpen certain textures. It's recommended to not use this model for very small images ( < 256px due to the large HR size it was trained on.

DDS Files with BC1/DXT1, BC3/DXT5 Compression

Name
Author
Scale
License
Purpose
Iterations
Batch Size
HR Size
Dataset Size
Dataset
Pretrained Model
"
bc1 sharp 3.0 Alsa 1 CC BY-NC-SA 4.0 BC1 Compression 400K 2 128 28.985K Custom (just about everything) BC1 take 2
bc1 smooth 2.0 BlueAmulet 1 CC BY-NC 4.0 BC1 Compression 1M 20 32 4767 Custom (? including normal maps) none (no interpolation)
BCGone_Smooth Mutin Choler 1 WTFPL BC1 Compression 110K 4 64 1400 Images from DIV2K and random anime drawings 1xESRGAN
BCGone_Detailed Mutin Choler 1 WTFPL BC1 Compression 115K 4 64 1400 Images from DIV2K and random anime drawings 1xESRGAN

BC1 (DXT1) compression is commonly used in dds textures, which are utilized in most PC games today, it allows to shrink the texture to 1/6 of the original size, reducing VRAM usage. There is also BC3 (DXT5) which uses BC1 compression for the color channels and leaves the alpha channel uncompressed, this one reduces the file size to 1/3 of the original. But this compression comes at a cost. The BC1 models are designed to reverse the damage done by the compression. This is a must if you want to upscale a dds file that uses either BC1 or BC3 compression. The free variant has more freedom when dealing with the images and should lead to better results. The restricted version only deals with perfect images compressed once (ideal case) so will perform worse in any other scenario, but it tries to preserve the original colors more. You can interpolate between restricted and unrestricted if you want.

Cinepak, msvideo1 and Roq

Name
Author
Scale
License
Purpose
Iterations
Batch Size
HR Size
Dataset Size
Dataset
Pretrained Model
"
Cinepak_alt buildist 1 GNU GPLv3 Cinepak (only) 240K 1 128 2,640 Flickr2K none (no interpolation)
Cinepak Twittman 1 CC BY-NC-SA 4.0 Cinepak, msvideo1 and Roq 200K 1 128 ~8K Custom (Manga) none (no interpolation)

The Cinepak model removes movie compressions artifacts from older video compression methods like Cinepak, msvideo1 and Roq. Cinepak_alt may work better for video frames with realistic content.

H264 Compression

Name
Author
Scale
License
Purpose
Iterations
Batch Size
HR Size
Dataset Size
Dataset
Pretrained Model
"
h264Texturize nmkd 1 WTFPL Heavy H264 Compression 500K 1 128 0.900K Unique DIV2K Train + Val 4xESRGAN

Indeo Compression Artifacts

Name
Author
Scale
License
Purpose
Iterations
Batch Size
HR Size
Dataset Size
Dataset
Pretrained Model
"
DeIndeo Wild West Quest#7975 4 WTFPL Indeo Compression Artifacts 130K 8 128 3.500K Custom 4xESRGAN

Color Reduction

Dithering / GIF

Name
Author
Scale
License
Purpose
Iterations
Batch Size
HR Size
Dataset Size
Dataset
Pretrained Model
"
dithering Alsa 1 CC BY-NC-SA 4.0 Dithered Images 127K 2 128 4.7K Custom (Photos / Manga) JPG (0-20%)
FSDedither buildist (Jacob_) 4 GNU GPLv3 Photos/realistic images with dithering 55K 16 128 ~20K Flickr2K+DIV2K dithered with GIMP 4xESRGAN
FSDedither Riven buildist (Jacob_) 4 GNU GPLv3 Photos/realistic images with dithering 200K 24 128 ~20K Flickr2K+DIV2K with different types of dithering applied 4xESRGAN
FSDedither Riven Smooth buildist (Jacob_) 4 GNU GPLv3 Photos/realistic images with dithering 191K 24 128 ~40K Flickr2K+DIV2K with different types of dithering applied 4xFSDedither_Riven
DeGif nmkd 2 WTFPL GIF Frames 160K 1 192 1.8K DIV2K Dithered + DIV2K Color Reduced 2xESRGAN
DitherDeleter Mutin Choler 1 WTFPL Dedithering 104K 4 64 300 Random anime images 1xESRGAN

Dithering is an older compression method, where the amount of colors gets reduced, if your image has few colors and havea noise like pattern in them try a De-Dither model. Ordered dithering is a less common form of dithering that results in distinctive checkerboard / crosshatch patterns, which are misinterpreted as texture by models not trained on it. It’s often used on GIFs because the pattern is stable between frames.

4sFSDedither is made for photos/realistic images, but worth trying on other images that have reduced colors and dithering along with fine details. Trained using the ESRGAN-FS code for better details compared to plain ESRGAN.

4xFSDedither_Riven is a remake of 4xFSDedither that is fine-tuned to upscale images from the game Riven, but should be better in general, particularly on ordered dithering. I adjusted the dataset to have a better variety of dithering parameters, and turned up the HFEN and pixel loss to get better details and color restoration with less noise. 4xFSDedither_Riven_Smooth adds images with nearest-neighbor downscaling to the dataset for a smoother (but possibly less detailed) result.

Banding

Name
Author
Scale
License
Purpose
Iterations
Batch Size
HR Size
Dataset Size
Dataset
Pretrained Model
"
N64clean BlueAmulet 1 CC BY-NC-SA 4.0 Cleaning N64 textures 277k 20 48 4034 (not tiled) CC0 Textures 1xESRGAN
Debandurh_FS Ultra-lite Twittman 1 CC BY-NC-SA 4.0 Debanding images 140k 1/4 128 14.4k Anime, real life, manga None

Banding is like dithering, but instead of adding patterns of pixels to simulate missing colors, it just clips them off.

Color Removal / Grayscale

Name
Author
Scale
License
Purpose
Iterations
Batch Size
HR Size
Dataset Size
Dataset
Pretrained Model
"
Colorizer BlackScout 1 GNU GPLv3 Coloring B/W images 22k 1 128 4.5045k Custom (800 Wallpapers) 1xESRGAN
SpongeColor Lite Joey 1 CC BY-NC-SA 4.0 Coloring B/W images 55k 12 64 24.91k Spongebob + DIV2K None

Just removes all colors and makes the image grayscale, the models here should restore some color to them, should also work for old black and white images.

Over-Sharpening

Name
Author
Scale
License
Purpose
Iterations
Batch Size
HR Size
Dataset Size
Dataset
Pretrained Model
"
DeSharpen loinne 1 CC BY-NC-SA 4.0 Oversharpened Images 310K 1 128 ~3K Custom (?) Failed Attempts

The De-Sharpen model was made for rare particular cases when the image was destroyed by applying noise, i.e. game textures or any badly exported photos. If your image does not have any over-sharpening, it won’t hurt them, leaving as is. In theory, this model knows when to activate and when to skip, also can successfully remove artifacts if only some parts of the image are over-sharpened, for example in image consisting of several combined images, 1 of them with sharpen noise. It is made to remove sharpen noise, particularly made with Photoshop “sharpen” or “sharpen more” filters OR ImageMagick’s -sharpen directive with several varying parameters of Radius and Sigma, from subtle 0.3x0.5 to something extreme like 8x2, somewhere about that.

Aliasing

Name
Author
Scale
License
Purpose
Iterations
Batch Size
HR Size
Dataset Size
Dataset
Pretrained Model
"
SS Anti Alias 9x BlueAmulet 1 CC BY-NC 4.0 Anti aliasing 125K~ 12 32 4096 Custom (9x Supersampling AA) none (no interpolation)
Anti Aliasing Twittman 1 CC BY-NC-SA 4.0 Images with pixelated edges 200K 1 128 0.656K Custom (?) none (no interpolation)

AntiAliasing is for smoothing jagged edges in images and textures.

Image Generation

Texture Maps

Name
Author
Scale
License
Purpose
Iterations
Batch Size
HR Size
Dataset Size
Dataset
Pretrained Model
"
normal generator LyonHrt 1 CC BY-NC-SA 4.0 Diffuse to Normal 215K 1 128 4.536K Custom (?) none (no interpolation)
NormalMapGenerator-CX Lite Joey 1 CC BY-NC-SA 4.0 Generating normal maps from textures 200k Varied 64 14.784K CC0 textures/normals, randomly downscaled OTF (by 1, 2, and 4) to get more variation DIV2K-Lite
FrankenMapGenerator-CX Lite Joey 1 CC BY-NC-SA 4.0 Generating roughness and displacement maps 215k Varied 64 14.784K CC0 textures/normals, randomly downscaled OTF (by 1, 2, and 4) to get more variation DIV2K-Lite

DeTooning

Name
Author
Scale
License
Purpose
Iterations
Batch Size
HR Size
Dataset Size
Dataset
Pretrained Model
"
Double Detoon Joey 1 CC BY-NC-SA 4.0 Detooning 100K 4 64 6.604K Reddit Images run through White-box-Cartoonization 1xESRGAN
Effeks Photoanime v2 Rek 4 GNU GPLv3 Photo stylization from JPEGs 150K 8 128 3.074K Frames extracted and upscaled from Hausu (1977) NMKD Yandere2

Pretrained models

Generator

Name Author Scale License Dataset Size Dataset Pretrained Model
1xESRGAN victorca25 1 Apache License 2.0 65k Combination of DIV2K, Flickr2K and GOPRO 4xPSNR
2xESRGAN victorca25 2 Apache License 2.0 65k Combination of DIV2K, Flickr2K and GOPRO 4xPSNR
4xESRGAN xinntao 4 Apache License 2.0 65k Combination of DIV2K, Flickr2K and GOPRO 4xPSNR
8xESRGAN victorca25 8 Apache License 2.0 65k Combination of DIV2K, Flickr2K and GOPRO 4xPSNR
16xESRGAN victorca25 16 Apache License 2.0 65k Combination of DIV2K, Flickr2K and GOPRO 4xPSNR
1xPSNR BlueAmulet 1 Apache License 2.0 65k Combination of DIV2K, Flickr2K and GOPRO none
2xPSNR BlueAmulet 2 Apache License 2.0 65k Combination of DIV2K, Flickr2K and GOPRO none
4xPSNR xinntao 4 Apache License 2.0 65k Combination of DIV2K, Flickr2K and GOPRO none
8xPSNR BlueAmulet 8 Apache License 2.0 65k Combination of DIV2K, Flickr2K and GOPRO none
16xPSNR BlueAmulet 16 Apache License 2.0 65k Combination of DIV2K, Flickr2K and GOPRO none

https://drive.google.com/drive/folders/1ldwajXL50uC7PCS63B4Wato6Dnk-svNL

These models were transformed from the original RRDB_ESRGAN_x4.pth model into the other scales, in order to be used as pretrained models for new models in those scales.

More information can be found here.

Discriminator

Name Author Scale License Dataset Size Dataset Pretrained Model
Pretrained Discriminator Pack Joey 1, 2, 4, 8 CC BY-NC-SA 4.0 ? ? none

SOFVSR ("vicGAN") Models

Real World

Name Author Scale License Purpose Iterations Batch Size HR Size Dataset Size Dataset Architecture Num Frames Pretrained Model
4x_REDSVAL-7f-RRDB-Lite Joey 4 CC BY-NC-SA 4.0 Upscaling real-world video / Pretrained model for 7-frames RRDB 200k Varied 128 3,000 REDS validation set RRDB 7 none

Cartoons

Name Author Scale License Purpose Iterations Batch Size HR Size Dataset Size Dataset Architecture Num Frames Pretrained Model
3x_Video_TSSM Joey 3 CC BY-NC-SA 4.0 Upscaling cartoon DVDs 70k 16 192 125,436 The Spongebob Squarepants Movie SOFVSR net 3 none
3x_Video_TSSM_RRDB Joey 3 CC BY-NC-SA 4.0 Upscaling cartoon DVDs 80k 6 192 125,436 The Spongebob Squarepants Movie RRDB 3 none
2x_SBS11-RRDB Joey 2 CC BY-NC-SA 4.0 Upscaling cartoon DVDs with bad haloing and noise 111k 6 128 126,956 Every 6 frames from every episode of Spongebob Season 11 (DVD and HD). DVD was resized with Spline36 to 540p. LR also got extra OTF JPEG (1/4 chance) and sharpening (15% chance) RRDB 3 none

Pretrained models

Name Author Scale License Purpose Iterations Batch Size HR Size Dataset Size Dataset Architecture Num Frames Pretrained Model
video_G Victorca25 4 ? Pretrained model for regular SOFVSR models ? 32 32 REDS' size REDS SOFVSR net 3 none
VESRGAN_G Victorca25 4 ? Pretrained model for RRDB SOFVSR models ? 32 32 REDS' size REDS RRDB 3 none