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
(32 intermediate revisions by 6 users not shown)
Line 35: Line 35:
 
! Purpose (Full)
 
! Purpose (Full)
 
! Pretrained_Model_G
 
! Pretrained_Model_G
! sample
+
! Sample
 
|-
 
|-
  
Line 48: Line 48:
 
|Streets with dense foliage in the background. Outdoor scenes.
 
|Streets with dense foliage in the background. Outdoor scenes.
 
|[https://drive.google.com/drive/folders/1ldwajXL50uC7PCS63B4Wato6Dnk-svNL 4xPSNR]
 
|[https://drive.google.com/drive/folders/1ldwajXL50uC7PCS63B4Wato6Dnk-svNL 4xPSNR]
|https://cdn.discordapp.com/attachments/547949806761410560/880291337373618236/unknown.png, https://media.discordapp.net/attachments/547949806761410560/880290211026853918/unknown.png, https://media.discordapp.net/attachments/547949806761410560/880299665428447232/unknown.png, https://cdn.discordapp.com/attachments/880291340301254657/880303552315146340/unknown.png
+
|[https://cdn.discordapp.com/attachments/547949806761410560/880291337373618236/unknown.png Sample 1]
 +
 
 +
[https://media.discordapp.net/attachments/547949806761410560/880290211026853918/unknown.png Sample 2]
 +
 
 +
[https://media.discordapp.net/attachments/547949806761410560/880299665428447232/unknown.png Sample 3]
 +
 
 +
[https://cdn.discordapp.com/attachments/880291340301254657/880303552315146340/unknown.png Sample 4]
 
|-
 
|-
  
Line 55: Line 61:
 
| 4x
 
| 4x
 
|[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]
|  
+
| ESRGAN
 
| General Upscaler
 
| General Upscaler
 
| 2021-04-09
 
| 2021-04-09
Line 72: Line 78:
 
|This is my best model yet! It generates lots and lots of detail and leaves a nice texture on images. It works on most images, whether compressed or not. It does work best on JPEG compression though, as that's mostly what it was trained on. It has the ability to restore highly compressed images as well! '''If you want a more balanced output, check out the UltraMix Collection down below. It's a bunch of interpolated models based around UltraSharp and my other models'''
 
|This is my best model yet! It generates lots and lots of detail and leaves a nice texture on images. It works on most images, whether compressed or not. It does work best on JPEG compression though, as that's mostly what it was trained on. It has the ability to restore highly compressed images as well! '''If you want a more balanced output, check out the UltraMix Collection down below. It's a bunch of interpolated models based around UltraSharp and my other models'''
 
|[https://mega.nz/#!uhAVEaAA!qfwIQ44Ba3rXMAb2-M3aM3zFBCwzmyQ64IO_0O-csJE 4xESRGAN]
 
|[https://mega.nz/#!uhAVEaAA!qfwIQ44Ba3rXMAb2-M3aM3zFBCwzmyQ64IO_0O-csJE 4xESRGAN]
|[https://cdn.discordapp.com/attachments/902853011913732097/902855337705611294/unknown.png Sample 1] [https://cdn.discordapp.com/attachments/902853011913732097/902853340600360990/unknown.png Sample 2] [https://cdn.discordapp.com/attachments/902853011913732097/902855036768489512/unknown.png Sample 3] [https://cdn.discordapp.com/attachments/900506271566938122/902867618086670356/unknown.png Sample 4]
+
|[https://cdn.discordapp.com/attachments/902853011913732097/902855337705611294/unknown.png Sample 1]
 +
 
 +
[https://cdn.discordapp.com/attachments/902853011913732097/902853340600360990/unknown.png Sample 2]
 +
 
 +
[https://cdn.discordapp.com/attachments/902853011913732097/902855036768489512/unknown.png Sample 3]
 +
 
 +
[https://cdn.discordapp.com/attachments/900506271566938122/902867618086670356/unknown.png Sample 4]
 
|-
 
|-
  
Line 108: Line 120:
 
|UniScale_Restore has strong compression removal that helps with restoring heavily compressed or noisy images. It is intended to compete with BSRGAN. Trained with BSRGAN_Resize and Combo_Noise in traiNNer.  
 
|UniScale_Restore has strong compression removal that helps with restoring heavily compressed or noisy images. It is intended to compete with BSRGAN. Trained with BSRGAN_Resize and Combo_Noise in traiNNer.  
 
|[https://mega.nz/#!uhAVEaAA!qfwIQ44Ba3rXMAb2-M3aM3zFBCwzmyQ64IO_0O-csJE 4xESRGAN]
 
|[https://mega.nz/#!uhAVEaAA!qfwIQ44Ba3rXMAb2-M3aM3zFBCwzmyQ64IO_0O-csJE 4xESRGAN]
|https://cdn.discordapp.com/attachments/880826637543964712/880934732748169226/unknown.png, https://cdn.discordapp.com/attachments/880826637543964712/880932754102026290/unknown.png
+
|[https://cdn.discordapp.com/attachments/880826637543964712/880934732748169226/unknown.png Sample 1]
 +
 
 +
[https://cdn.discordapp.com/attachments/880826637543964712/880932754102026290/unknown.png Sample 2]
 
|-
 
|-
  
Line 120: Line 134:
 
| These models work great on game textures when interpolated 50/50 with UniScale_Restore, and work amazingly on uncompressed images. DO NOT USE FOR COMPRESSED IMAGES, use the original UniScale or UltraSharp for that.
 
| These models work great on game textures when interpolated 50/50 with UniScale_Restore, and work amazingly on uncompressed images. DO NOT USE FOR COMPRESSED IMAGES, use the original UniScale or UltraSharp for that.
 
|[https://mega.nz/#!uhAVEaAA!qfwIQ44Ba3rXMAb2-M3aM3zFBCwzmyQ64IO_0O-csJE 4xESRGAN]
 
|[https://mega.nz/#!uhAVEaAA!qfwIQ44Ba3rXMAb2-M3aM3zFBCwzmyQ64IO_0O-csJE 4xESRGAN]
|https://cdn.discordapp.com/attachments/884239326471393331/884296266857713704/unknown.png https://cdn.discordapp.com/attachments/884239326471393331/884294468709257216/unknown.png https://cdn.discordapp.com/attachments/884239326471393331/884298572894441512/unknown.png https://cdn.discordapp.com/attachments/884239326471393331/884302370136289323/unknown.png
+
|[https://cdn.discordapp.com/attachments/884239326471393331/884296266857713704/unknown.png Sample 1]
 +
 
 +
[https://cdn.discordapp.com/attachments/884239326471393331/884294468709257216/unknown.png Sample 2]
 +
 
 +
[https://cdn.discordapp.com/attachments/884239326471393331/884298572894441512/unknown.png Sample 3]
 +
 
 +
[https://cdn.discordapp.com/attachments/884239326471393331/884302370136289323/unknown.png Sample 4]
 
|-
 
|-
  
Line 211: Line 231:
 
! Purpose (Full)
 
! Purpose (Full)
 
! Pretrained_Model_G
 
! Pretrained_Model_G
! sample
+
! Sample
  
 
|-
 
|-
Line 223: Line 243:
 
| This model was made to upscale realistic low-res textures that are compressed by either JPEG or BC1. From my testing, this works rather well on realistic GameCube textures such as the ones from Shrek Extra Large and the board textures from Mario Party 4. This model could also work on some real life images, especially the ones that are taken outdoors.
 
| This model was made to upscale realistic low-res textures that are compressed by either JPEG or BC1. From my testing, this works rather well on realistic GameCube textures such as the ones from Shrek Extra Large and the board textures from Mario Party 4. This model could also work on some real life images, especially the ones that are taken outdoors.
 
| RealESRGAN_x4plus
 
| RealESRGAN_x4plus
| https://imgsli.com/MTQ0Mzk1 https://imgsli.com/MTQ0MTYy https://imgsli.com/MTQ0MjEx https://imgsli.com/MTQ0MTQ2 https://imgsli.com/MTQ0MTQ3 https://imgsli.com/MTQ0MTQ5 https://imgsli.com/MTQ0MTUw https://imgsli.com/MTQ0MTUx https://imgsli.com/MTQ0MTUz https://imgsli.com/MTQ0MTU1 https://imgsli.com/MTQ0MTU2 https://imgsli.com/MTQ0MTU4 https://imgsli.com/MTQ0MTU5 https://imgsli.com/MTQ0MTYw https://imgsli.com/MTQ0MTYx https://imgsli.com/MTQ0MjEz
+
| https://imgsli.com/MTQ0Mzk1
 +
 
 +
https://imgsli.com/MTQ0MTYy
 +
 
 +
https://imgsli.com/MTQ0MjEx
 +
 
 +
https://imgsli.com/MTQ0MTQ2
 +
 
 +
https://imgsli.com/MTQ0MTQ3
 +
 
 +
https://imgsli.com/MTQ0MTQ5
 +
 
 +
https://imgsli.com/MTQ0MTUw
 +
 
 +
https://imgsli.com/MTQ0MTUx
 +
 
 +
https://imgsli.com/MTQ0MTUz
 +
 
 +
https://imgsli.com/MTQ0MTU1
 +
 
 +
https://imgsli.com/MTQ0MTU2
 +
 
 +
https://imgsli.com/MTQ0MTU4
 +
 
 +
https://imgsli.com/MTQ0MTU5
 +
 
 +
https://imgsli.com/MTQ0MTYw
 +
 
 +
https://imgsli.com/MTQ0MTYx
 +
 
 +
https://imgsli.com/MTQ0MjEz
 
|-
 
|-
  
Line 312: Line 362:
 
! Purpose (Full)
 
! Purpose (Full)
 
! Pretrained_Model_G
 
! Pretrained_Model_G
! sample
+
! Sample
  
 
|-
 
|-
Line 390: Line 440:
  
  
| [https://1drv.ms/u/s!Aip-EMByJHY20xsc9EvnxVZ50Ja7 Lady0101_208000.pth]
+
| [https://1drv.ms/u/s!Aip-EMByJHY20xsc9EvnxVZ50Ja7?e=Loa5Gz Lady0101_208000.pth]
| DinnJerr
+
| [[User:DinJerr|DinJerr]]
 
| 4x
 
| 4x
 
|  
 
|  
Line 465: Line 515:
 
| Similar to Manga109, can be used as a general digital upscaler as well as with pixel art
 
| Similar to Manga109, can be used as a general digital upscaler as well as with pixel art
 
| RRDB_PSNR_x4
 
| RRDB_PSNR_x4
| https://www.youtube.com/channel/UCwBfuiHdSPQ-zslOmiW8OHg
+
| [https://www.youtube.com/channel/UCwBfuiHdSPQ-zslOmiW8OHg Video Samples]
 
|-
 
|-
  
Line 494: Line 544:
  
 
| [https://1drv.ms/u/s!Aip-EMByJHY26nuB0Lo0rdirfGoi?e=XI9Fa2 8x_HugePeeps_v1]
 
| [https://1drv.ms/u/s!Aip-EMByJHY26nuB0Lo0rdirfGoi?e=XI9Fa2 8x_HugePeeps_v1]
| DinJerr
+
| [[User:DinJerr|DinJerr]]
 
| 8x
 
| 8x
 
|[https://creativecommons.org/licenses/by-nc/4.0/ CC BY-NC 4.0]
 
|[https://creativecommons.org/licenses/by-nc/4.0/ CC BY-NC 4.0]
Line 505: Line 555:
 
|-
 
|-
  
| [https://1drv.ms/u/s!Aip-EMByJHY22BhfdJ1Epkfc3ugq?e=yVfXkz ArtStation]
+
| [https://1drv.ms/u/s!Aip-EMByJHY21WF47rcqn7fhr7fq?e=I8mnX6 ArtStation1337]
| DinJerr
+
| [[User:DinJerr|DinJerr]]
| 8x
+
| 4x
| ?
+
|[https://creativecommons.org/licenses/by-nc/4.0/ CC BY-NC 4.0]
 
| ESRGAN
 
| ESRGAN
| Art/People
+
| Digital Art/People
 
| 2019-08-21
 
| 2019-08-21
 
|  
 
|  
| '''GO UP A FOLDER, THEN TO THE ARTSTAION1000 OR ARTSTATION1337 FOLDER'''
+
| Mainly for digital art, but can be used to upscale pixel art.
 
|  
 
|  
 
|-
 
|-
Line 551: Line 601:
 
| Upscaling airbrush/pencil-based artwork
 
| Upscaling airbrush/pencil-based artwork
 
| 4xPSNR
 
| 4xPSNR
| https://drive.google.com/drive/folders/1fyeIWInDrM6r-xxrCW4U09oafWw08u9S?usp=sharing
+
| [https://drive.google.com/drive/folders/1fyeIWInDrM6r-xxrCW4U09oafWw08u9S?usp=sharing Sample Gallery]
 
|-
 
|-
  
Line 567: Line 617:
 
|-
 
|-
  
| [https://1drv.ms/u/s!Aip-EMByJHY200v-vXPezsCsUd5T?e=r7mBNn BigFArt]
+
| [https://1drv.ms/u/s!Aip-EMByJHY23if0ac150Sj2qnbp?e=qLp7SP BigFArt]
| DinJerr
+
| [[User:DinJerr|DinJerr]]
 
| 4x
 
| 4x
 
|[https://creativecommons.org/licenses/by-nc/4.0/ CC BY-NC 4.0]
 
|[https://creativecommons.org/licenses/by-nc/4.0/ CC BY-NC 4.0]
Line 604: Line 654:
 
|-
 
|-
  
| [https://1drv.ms/u/s!Aip-EMByJHY22BhfdJ1Epkfc3ugq?e=yVfXkz UltraFArt_v3 / FArtDIV3 Suite]
+
| [https://1drv.ms/u/s!Aip-EMByJHY27UIFr1mk87zFytap?e=SSxKaV UltraFArt_v3 Suite]
| DinJerr
+
| [[User:DinJerr|DinJerr]]
 
| 4x
 
| 4x
 
|[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]
Line 611: Line 661:
 
| Art
 
| Art
 
| 2021-05-14
 
| 2021-05-14
| Painting style with larger shaped features (?).
+
| Illustrations with with larger shaped features (?).
 
| 4x_UltraFArt
 
| 4x_UltraFArt
| https://1drv.ms/u/s!Aip-EMByJHY27UIFr1mk87zFytap?e=ZZliII
+
|  
 
|-
 
|-
  
| [https://drive.google.com/drive/folders/1mYTMpwDlKQulBmjgKpcxL3-T6xAGXVxl?usp=sharing NXbrz]
+
|-
| Archerpolation
+
| [https://1drv.ms/u/s!Aip-EMByJHY26wU-wOqApp7Hu7tt?e=4ihnBr HugePaint]
| 4x
+
| |[[User:DinJerr|DinJerr]]
 +
| 8x
 +
| [https://creativecommons.org/licenses/by-nc/4.0/ CC BY-NC 4.0]
 +
| ESRGAN
 +
| Digital Illustrations
 +
|
 +
| Trained on a variety of images from ArtStation
 +
| HugePeeps
 +
|
 +
|-
 +
 
 +
| [https://drive.google.com/drive/folders/1mYTMpwDlKQulBmjgKpcxL3-T6xAGXVxl?usp=sharing NXbrz]
 +
| Archerpolation
 +
| 4x
 
|[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]
 
| ESRGAN
 
| ESRGAN
Line 624: Line 687:
 
| 2021-06-09
 
| 2021-06-09
 
| Basic pixel art upscaling, for people who want a more simpler style and lightweight pixel art upscaling model.
 
| Basic pixel art upscaling, for people who want a more simpler style and lightweight pixel art upscaling model.
| https://drive.google.com/drive/folders/1mYTMpwDlKQulBmjgKpcxL3-T6xAGXVxl?usp=sharing
+
|
 +
| [https://drive.google.com/drive/folders/1mYTMpwDlKQulBmjgKpcxL3-T6xAGXVxl?usp=sharing Sample Gallery]
 
|-
 
|-
 
|}
 
|}
Line 630: Line 694:
 
===Drawn Material===
 
===Drawn Material===
  
====Anime====
+
====Anime and Cartoons====
  
 
{| class="wikitable sortable"
 
{| class="wikitable sortable"
Line 643: Line 707:
 
! Purpose (Full)
 
! Purpose (Full)
 
! Pretrained_Model_G
 
! Pretrained_Model_G
! sample
+
! Sample
 
 
  
 
|-
 
|-
Line 656: Line 719:
 
| Interpolation between 4x-UltraSharp and 4x-TextSharp-v0.5. Works amazingly on anime. It also upscales text, but it's far better with anime content. I rebranded this model on 2/10/22 to 4x-AnimeSharp from 4x-TextSharpV1.
 
| Interpolation between 4x-UltraSharp and 4x-TextSharp-v0.5. Works amazingly on anime. It also upscales text, but it's far better with anime content. I rebranded this model on 2/10/22 to 4x-AnimeSharp from 4x-TextSharpV1.
 
|  
 
|  
| [https://cdn.discordapp.com/attachments/903415274521374750/925533775616696340/unknown.png Text Sample] [https://cdn.discordapp.com/attachments/549525506585001985/941432207917084743/Clipboard-comparison.png Anime Sample] [https://cdn.discordapp.com/attachments/549525506585001985/941432638206541825/test_pic-comparison.png Anime Sample 2]
+
| [https://cdn.discordapp.com/attachments/903415274521374750/925533775616696340/unknown.png Text Sample]  
 +
 
 +
[https://cdn.discordapp.com/attachments/549525506585001985/941432207917084743/Clipboard-comparison.png Anime Sample]  
 +
 
 +
[https://cdn.discordapp.com/attachments/549525506585001985/941432638206541825/test_pic-comparison.png Anime Sample 2]  
  
 
|-
 
|-
Line 668: Line 735:
 
| This model is a lite version of AnimeSharp. It was trained using student-teacher learning (if i'm using the term properly), where the HRs are LRs upscaled by the full size AnimeSharp ESRGAN model, and the lite model is trained on those outputs as the HR. It works best on clean or slightly blurry anime. '''Downscale by 50% first in almost all cases'''
 
| This model is a lite version of AnimeSharp. It was trained using student-teacher learning (if i'm using the term properly), where the HRs are LRs upscaled by the full size AnimeSharp ESRGAN model, and the lite model is trained on those outputs as the HR. It works best on clean or slightly blurry anime. '''Downscale by 50% first in almost all cases'''
 
|  
 
|  
| [https://cdn.discordapp.com/attachments/547949405949657100/941776512049360986/unknown.png Sample 1] [https://cdn.discordapp.com/attachments/547949405949657100/941770824543797268/unknown.png Sample 2] [https://cdn.discordapp.com/attachments/903415274521374750/941779970013925506/unknown.png Sample 3]
+
| [https://cdn.discordapp.com/attachments/547949405949657100/941776512049360986/unknown.png Sample 1]  
 +
 
 +
[https://cdn.discordapp.com/attachments/547949405949657100/941770824543797268/unknown.png Sample 2]  
 +
 
 +
[https://cdn.discordapp.com/attachments/903415274521374750/941779970013925506/unknown.png Sample 3]  
  
 
|-
 
|-
| [https://e1.pcloud.link/publink/show?code=kZ7rGRZW2IcOpNMQeXDTTRQ4aPVBFyyJV5X sudo_RealESRGAN2x_3.332.758_G.pth / sudo_RealESRGAN2x_Dropout_3.799.042_G.pth (pcloud)] [https://www.mediafire.com/folder/gk7f61e2kut0z/sudo_RealESRGAN2x sudo_RealESRGAN2x_3.332.758_G.pth / sudo_RealESRGAN2x_Dropout_3.799.042_G.pth (mediafire)]
+
| <small>[https://e1.pcloud.link/publink/show?code=kZ7rGRZW2IcOpNMQeXDTTRQ4aPVBFyyJV5X sudo_RealESRGAN2x_3.332.758_G.pth / sudo_RealESRGAN2x_Dropout_ 3.799.042_G.pth (pcloud)]</small>
 +
<small>[https://www.mediafire.com/folder/gk7f61e2kut0z/sudo_RealESRGAN2x sudo_RealESRGAN2x_3.332.758_G.pth / sudo_RealESRGAN2x_Dropout_ 3.799.042_G.pth (mediafire)]</small>
 
| sudo
 
| sudo
 
| 2x
 
| 2x
Line 679: Line 751:
 
| 2022-06-25
 
| 2022-06-25
 
| Tried to make the best 2x model there is for drawings. I think i archived that. And yes, it is nearly 3.8 million iterations (probably a record nobody will beat here), took me nearly half a year to train. It can happen that in one edge is a noisy pattern in edges. You can use padding/crop for that. I aimed for perceptual quality without zooming in like 400%. Since RealESRGAN is 4x, I downscaled these images with bicubic. I would recommend my VSGAN code though and just load the onnx. https://github.com/styler00dollar/VSGAN-tensorrt-docker I just wanted a good 2x model for animations, but that model can also be used for wallpapers and so on. Before I hear people complaining, the dropout model is a modified architecture. Stuff like cupscale or chaiNNer won't work with pth. Load the onnx with VSGAN or chaiNNer. I did add the model before switching to dropout though, which is normal ESRGAN pth, that one should work everywhere. I also converted everything into onnx, jit and ncnn, so pretty much everything there is. If you want to use ncnn, don't use nihuis code (that also includes cupscale), these codes don't include propper tiling in C++, which is very bad for this model. I think chaiNNer should have overlap/padding with ncnn, so use that instead if you really want ncnn. Plz don't steal without credits, k thx.
 
| Tried to make the best 2x model there is for drawings. I think i archived that. And yes, it is nearly 3.8 million iterations (probably a record nobody will beat here), took me nearly half a year to train. It can happen that in one edge is a noisy pattern in edges. You can use padding/crop for that. I aimed for perceptual quality without zooming in like 400%. Since RealESRGAN is 4x, I downscaled these images with bicubic. I would recommend my VSGAN code though and just load the onnx. https://github.com/styler00dollar/VSGAN-tensorrt-docker I just wanted a good 2x model for animations, but that model can also be used for wallpapers and so on. Before I hear people complaining, the dropout model is a modified architecture. Stuff like cupscale or chaiNNer won't work with pth. Load the onnx with VSGAN or chaiNNer. I did add the model before switching to dropout though, which is normal ESRGAN pth, that one should work everywhere. I also converted everything into onnx, jit and ncnn, so pretty much everything there is. If you want to use ncnn, don't use nihuis code (that also includes cupscale), these codes don't include propper tiling in C++, which is very bad for this model. I think chaiNNer should have overlap/padding with ncnn, so use that instead if you really want ncnn. Plz don't steal without credits, k thx.
| Pretrained_Model_G: RealESRGAN_x4plus_anime_6B.pth / RealESRGAN_x4plus_anime_6B.pth (sudo_RealESRGAN2x_3.332.758_G.pth)
+
| Pretrained_Model_G: RealESRGAN_x4plus_anime_6B.pth /
| [https://cdn.discordapp.com/attachments/579685650824036387/990323693660033045/jpg_compare.png Sample 1] [https://cdn.discordapp.com/attachments/579685650824036387/990323694196916275/test2_compare.png Sample 2] [https://cdn.discordapp.com/attachments/579685650824036387/990323694561816606/f4_compare.png Sample 3] [https://cdn.discordapp.com/attachments/579685650824036387/990323695186763884/test_compare.png Sample 4][https://cdn.discordapp.com/attachments/579685650824036387/990323695601987584/f2_compare.png Sample 5]
+
RealESRGAN_x4plus_anime_6B.pth  
 +
(sudo_RealESRGAN2x_ 3.332.758_G.pth)
 +
| [https://cdn.discordapp.com/attachments/579685650824036387/990323693660033045/jpg_compare.png Sample 1]
 +
 
 +
[https://cdn.discordapp.com/attachments/579685650824036387/990323694196916275/test2_compare.png Sample 2]
 +
 
 +
[https://cdn.discordapp.com/attachments/579685650824036387/990323694561816606/f4_compare.png Sample 3]
 +
 
 +
[https://cdn.discordapp.com/attachments/579685650824036387/990323695186763884/test_compare.png Sample 4]
 +
 
 +
[https://cdn.discordapp.com/attachments/579685650824036387/990323695601987584/f2_compare.png Sample 5]
 +
|-
 +
 
 +
| <small>[https://e1.pcloud.link/publink/show?code=kZufYRZj1INMisvFehhfS763L4ow5YUy0VV sudo_UltraCompact_2x_1.121.175_G.pth (pcloud)] / [https://www.mediafire.com/folder/7e7752gf42eky/sudo_UltraCompact_2x_1.121.175 sudo_UltraCompact_2x_1.121.175_G.pth (mediafire)]</small>
 +
| sudo
 +
| 2x
 +
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 +
| Compact
 +
| Realtime animation restauration and doing stuff like deblur and compression artefact removal
 +
| 2022-05-29
 +
| My first attempt to make a REALTIME 2x upscaling model while also applying teacher student learning. It beats Anime4k in every way. These benchmarks use a 3060ti and it shows that everything better than a 3060ti should be able to handle 1080p input if you create engine files and use my TensorRT code.  You can see in the readme how to convert onnx files into engines. The 2 right bars compare normal Compact2 and Ultracompact in speed, the 2 on the left showcase older apis I used which isn't too important for this showcase.  To use this, you need to use my code which is https://github.com/styler00dollar/VSGAN-tensorrt-docker. If you use Manjaro, it is also possible to pipe the data stream directly into mpv, so you can watch it in a video player without rendering a video. Yeah the model does seem a little noisy if you zoom in a lot, but don't forget that the model itself is only 1.2mb. I think it does quite well. I still try to improve on fast models, but this is good enough to share as a first model. Plz don't steal without credits, k thx.
 +
| RealESRGANv2-animevideo-xsx2.pth
 +
(Teacher: RealESRGANv2-animevideo-xsx2.pth)
 +
| [https://cdn.discordapp.com/attachments/579685650824036387/980550564725260358/compare.png Sample 1]
 +
 
 +
[https://cdn.discordapp.com/attachments/579685650824036387/980550565354410066/unknown.png Sample 2]
 +
 
 +
[https://cdn.discordapp.com/attachments/579685650824036387/980550565723529226/compare2.webp Sample 3]
 
|-
 
|-
  
 
|-
 
|-
| [https://github.com/Bubblemint864/AI-Models/releases/tag/2x_Bubble_AnimeScale_Compact_v1 2x_Bubble_AnimeScale_Compact_v1]
+
| <small>[https://github.com/Bubblemint864/AI-Models/releases/tag/2x_Bubble_AnimeScale_Compact_v1 2x_Bubble_AnimeScale_Compact_v1]</small>
 
| Bubblemint#6472
 
| Bubblemint#6472
 
| 2x
 
| 2x
Line 694: Line 793:
 
| 4x_muy4_035_1.pth
 
| 4x_muy4_035_1.pth
 
| https://imgsli.com/MTM0MzMx  
 
| https://imgsli.com/MTM0MzMx  
 +
 
https://imgsli.com/MTM0MzM3  
 
https://imgsli.com/MTM0MzM3  
 +
 
https://imgsli.com/MTM0MzQw
 
https://imgsli.com/MTM0MzQw
  
 
|-
 
|-
| [https://github.com/Bubblemint864/AI-Models/releases/tag/2x_Bubble_AnimeScale_SwinIR_Small_v1 2x_Bubble_AnimeScale_SwinIR_Small_v1]
+
| <small>[https://github.com/Bubblemint864/AI-Models/releases/tag/2x_Bubble_AnimeScale_SwinIR_Small_v1 2x_Bubble_AnimeScale_SwinIR_Small_ v1]</small>
 
| Bubblemint#6472
 
| Bubblemint#6472
 
| 2x
 
| 2x
Line 705: Line 806:
 
| Anime or Text
 
| Anime or Text
 
| 2022-11-13
 
| 2022-11-13
| 2x_Bubble_AnimeScale_SwinIR_Small_v1 was trained to upscale anime frames faithfully without major contrast shifting compared to my compact model. Although much slower compared to my compact model, the results look significantly better! A few example upscales are listed below; more can be found by clicking the Overview link on the Github release page.
+
| 2x_Bubble_AnimeScale_SwinIR_Small_ v1 was trained to upscale anime frames faithfully without major contrast shifting compared to my compact model. Although much slower compared to my compact model, the results look significantly better! A few example upscales are listed below; more can be found by clicking the Overview link on the Github release page.
 
| None
 
| None
 
| https://imgsli.com/MTM2MjAx  
 
| https://imgsli.com/MTM2MjAx  
 +
 
https://imgsli.com/MTM2MjAy  
 
https://imgsli.com/MTM2MjAy  
 +
 
https://imgsli.com/MTM2MjAz
 
https://imgsli.com/MTM2MjAz
  
Line 719: Line 822:
 
| Anime / Visual Novel Art
 
| Anime / Visual Novel Art
 
| 2022-12-17
 
| 2022-12-17
| Third iteration of my eroge upscaling model. Discriminator: https://drive.google.com/file/d/18q-4ktFNZ8tPjkszuoG6kVBhCjKl8iFa/view?usp=sharing
+
| Third iteration of my eroge upscaling model. Discriminator: [https://drive.google.com/file/d/18q-4ktFNZ8tPjkszuoG6kVBhCjKl8iFa/view?usp=sharing Google Drive]
 
|  
 
|  
 
| https://slow.pics/c/fqjAhnjH
 
| https://slow.pics/c/fqjAhnjH
Line 733: Line 836:
 
| Upscaling cel animation. Trained this more than a year ago, releasing cause I've got a much better v2 and v3 now.
 
| Upscaling cel animation. Trained this more than a year ago, releasing cause I've got a much better v2 and v3 now.
 
| 4xPSNR
 
| 4xPSNR
| https://slow.pics/c/UrnFYuXX https://slow.pics/c/zy1Kd841
+
| https://slow.pics/c/UrnFYuXX
 +
 
 +
https://slow.pics/c/zy1Kd841
  
 
|-
 
|-
| [https://drive.google.com/file/d/1bwJpBHynGTIedIaZwp5XapCrJjO4BruF/view?usp=share_link 2xGT-evA.pth]
+
| [https://mega.nz/folder/XZgSmAoa#KNKbXZVBDq4UD1NJLHm1oQ 2x / 4x-anifilm_compact]
| evA-01
+
| [[User:Kim2091|Kim2091]]
| 2x
+
| 2x and 4x
 
| CC BY-NC-SA 4.0
 
| CC BY-NC-SA 4.0
| Real-ESRGAN Compact
+
| Compact
| Anime or Text
+
| Animation
| 2022-12-14
+
| 2022-08-02
| This is my first 2xcompact model, the main purpose of it is to upscale Dragon ball GT. for upscaling videos that have grain I would recommend denoising and dehaloing it before passing it to the model for temporal stability.
+
| This model is based on a private model by @eula 5600x 3070 named 4x_eula_anifilm_v1_225k. He sent me a copy of the model, and I decided to train a compact model based on it with his permission. This model seems to fix the majority of the issues the original model had while being far faster, it's just a tiny bit softer in some images.
| 2x_Compact_Pretrain.pth
+
 
| [https://cdn.discordapp.com/attachments/579685650824036387/1051753650051088384/test_6.png Sample 1] [https://cdn.discordapp.com/attachments/579685650824036387/1051753650428588092/test_2.png Sample 2] [https://cdn.discordapp.com/attachments/579685650824036387/1051753651011600444/test3.png Sample 3] [https://cdn.discordapp.com/attachments/579685650824036387/1051753651376504842/test_3.png Sample 4]
+
The dataset consists of Dragon Ball movies converted to YUV24 with @sgdisk --zap-all /dev/sda's help to reduce artifacts, then upscaled with ArtClarity and eula_anifilm. LRs are the original frames right from DVD. As a result, this model corrects some color space issues. The 2x model's HRs were downscaled by 50% with Lanczos.
 +
 
 +
The 2x and 4x models are pretty close in output despite being trained separately. The 2x model is a bit softer overall.
 +
 
 +
The models in the Real-ESRGAN Compatible folder are the original output from Real-ESRGANs training code for compatibility reasons.
 +
| 4x_Compact_Pretrain.pth
 +
| 2x Comparison:
 +
 
 +
[https://cdn.discordapp.com/attachments/903415274521374750/1004152057739087972/1659478894.4425747.png Sample 1]
 +
 
 +
[https://imgsli.com/MTE5MjUz Slider Comparison 1]
 +
 
 +
4x Comparison:  
  
|-
+
[https://imgsli.com/MTE5MjU0 Sample 2]
| [https://drive.google.com/drive/folders/13PhLXD3Pvo1A7LwHGIeY_ekObva_333e?usp=share_link 1xEpsilon-one-compact.pth]
 
| evA-01
 
| 1x
 
| CC BY-NC-SA 4.0
 
| Real-ESRGAN Compact
 
| Anime or Text
 
| 2022-12-03
 
| This model is far from perfect, but it does a decent job on removing dot crawl and dehalo\deblock old anime at fast rate without removing many details. chaining it with models like 2x-anifilm-compact or 2xLD-ANIME or 2x_AnimeClassics_UltraLite_510K will give good results i think.
 
| 1x_Compact_Pretrain.pth
 
| https://imgsli.com/MTM3Mzgz/2/3
 
https://imgsli.com/MTM3Mzg1/2/3
 
https://imgsli.com/MTM3Mzg2/0/1
 
  
 +
[https://cdn.discordapp.com/attachments/903415274521374750/1004158991544365097/1659480539.543144.png Slider Comparison 2]
  
 
|-
 
|-
| [https://drive.google.com/open?id=1H3F8OVBnK2cd5NjbCyc3tGwTZJBF1gk7 Sol Levante NTSC2HD]
+
| [https://mega.nz/file/xI8D2SCJ#oQ0h3k9lemeNJcuysJql7z55WV_NhOsKRpzpErIhcGc LD-Anime_Compact]
| Phoenix
+
| Zarxrax / Skr
| 4x
+
| 2x
| UNLICENSE
+
| [https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
| ESRGAN
+
| Compact
| Anime/Pretrained
+
| Animation
| 2020-04-08
+
| 2022-12-22
| NTSC DVD-spec encode x4 scale super-resolution for Anime Drawing style content. The dataset has a LOT of data throughout almost every frame, so it had a lot of stuff to learn. The resulting DVD-spec encode also had some blocking at times so it also learned to fight off blocking.
+
| I trained Skr's great LD-Anime model on compact architecture. It upscales while fixing numerous video problems, including: noise/grain, compression artifacts, rainbows, dot crawl, halos and color bleed. This compact version may look slightly worse than Skr's original model, but runs significantly faster and also retains the correct colors better than the original model did.
| RRDB_PSNR_x4
+
| 2x_Compact_Pretrain.pth
|  
+
| https://imgsli.com/MTQyMzM3/0/1
 +
 
 
|-
 
|-
 +
| [https://mega.nz/file/wdszCQRZ#bs7VIgvjgRiKiL5JVPqzCLlNl0kIQnDeD48-6ltWSLw Futsuu Anime]
 +
| Zarxrax
 +
| 2x
 +
| WTFPL
 +
| Compact
 +
| Animation
 +
| 2023-1-18
 +
| This model upscales while doing some sharpening and line darkening. Can also clean up some minor artifacts of various types. It is intended to to be a good general purpose upscaler that will work well with most animation.
 +
| 2x_Compact_Pretrain.pth
 +
| https://imgsli.com/MTQ4MDM2/
  
 
|-
 
|-
| [https://mega.nz/file/taBFxAAR#6wW-DrGh6oy2uz6aAKJpa36ngE2eKcDNA8Psy11oK10 2x_AnimeClassics_UltraLite_510K]
+
| [https://drive.google.com/file/d/1bwJpBHynGTIedIaZwp5XapCrJjO4BruF/view?usp=share_link 2xGT-evA.pth]
| CG1989
+
| evA-01
 
| 2x
 
| 2x
 
| CC BY-NC-SA 4.0
 
| CC BY-NC-SA 4.0
| ESRGAN
+
| Real-ESRGAN Compact
| Anime/Pretrained
+
| Anime or Text
| 2022-03-08
+
| 2022-12-14
| A 2x Ultra Lite model coming in under 8MB. Trained with over 15 sets of LRs ranging in a wide amount of issues. Handles Rainbows, Dot Crawl, MPEG/H.264 Compression, and may even assist in removing halos, and fixing blurriness in certain cases. This is my first public model for everyone. Best when used on old anime that is grainy. I can't say what anime it's best suited for as I have tried multiple series, and have found it does a good job on most all the tests. I wouldn't say use this for Western Animation, but it may work. I have done a few tests that I have shown in the upscale results, but that was chained with other models to achieve such a result. This model is meant to retain the more natural look of a series. There is a color shift on the end result, not drastic, but still noticable. I figure you should fix any color issues in post that way to give a more polished upscale. Big thanks to @SaurusX for the model name, and just helping out in general with anything.
+
| This is my first 2xcompact model, the main purpose of it is to upscale Dragon ball GT. for upscaling videos that have grain I would recommend denoising and dehaloing it before passing it to the model for temporal stability.
| 2X_DigitalFilmV5_Lite.pth
+
| 2x_Compact_Pretrain.pth
| [https://imgsli.com/OTg3MjM Sample 1] [https://imgsli.com/OTg3MjQ Sample 2] [https://imgsli.com/OTg3MjU Sample 3] [https://imgsli.com/OTg3Nzg Sample 4]
+
| [https://cdn.discordapp.com/attachments/579685650824036387/1051753650051088384/test_6.png Sample 1]
 +
 
 +
[https://cdn.discordapp.com/attachments/579685650824036387/1051753650428588092/test_2.png Sample 2]
 +
 
 +
[https://cdn.discordapp.com/attachments/579685650824036387/1051753651011600444/test3.png Sample 3]
  
|-
+
[https://cdn.discordapp.com/attachments/579685650824036387/1051753651376504842/test_3.png Sample 4]
  
 
|-
 
|-
| [https://drive.google.com/file/d/1DRyqqy24OU6G_mwsGjy6Zaxan_Fy1LNU/view?usp=sharing 600k] [https://drive.google.com/file/d/1psKp2DLWscPmDuzSTC8XqXL0LWnWuTYs/view?usp=sharing 650k] BooruGan
+
| [https://drive.google.com/drive/folders/13PhLXD3Pvo1A7LwHGIeY_ekObva_333e?usp=share_link 1xEpsilon-one-compact.pth]
| Tal
+
| evA-01
| 4x
+
| 1x
| WTFPL
+
| CC BY-NC-SA 4.0
| ESRGAN
+
| Real-ESRGAN Compact
| Anime
+
| Anime or Text
| 2021-11-26
+
| 2022-12-03
| This model is designed to mainly upscale anime artworks. If you have issues with chroma then try the 600k iterations release.
+
| This model is far from perfect, but it does a decent job on removing dot crawl and dehalo\deblock old anime at fast rate without removing many details. chaining it with models like 2x-anifilm-compact or 2xLD-ANIME or 2x_AnimeClassics_UltraLite_510K will give good results i think.
| 4x_Manga109Attempt
+
| 1x_Compact_Pretrain.pth
| [https://cdn.discordapp.com/attachments/579685650824036387/913500929012150323/4f012e36d686033b274261295da89b32-comparison.jpg Sample 1] [https://cdn.discordapp.com/attachments/579685650824036387/913500929356070912/256x256fdbb-comparison.jpg Sample 2] [https://cdn.discordapp.com/attachments/579685650824036387/913500929649688617/franky-super-810x456-comparison.jpg Sample 3]
+
| https://imgsli.com/MTM3Mzgz/2/3
|-
+
 
 +
https://imgsli.com/MTM3Mzg1/2/3
 +
 
 +
https://imgsli.com/MTM3Mzg2/0/1
 +
 
  
 
|-
 
|-
| [https://drive.google.com/file/d/1FWzEgqaSg2z0qOSjcmAxEUB3clEkm1PG/view 2x_Byousoku_5_Centimeter.pth]
+
| [https://drive.google.com/open?id=1H3F8OVBnK2cd5NjbCyc3tGwTZJBF1gk7 Sol Levante NTSC2HD]
| Mystery_Bullet#0642 (Shady Adel)
+
| Phoenix
| 2x
+
| 4x
| [https://www.gnu.org/licenses/gpl-3.0.en.html GNU GPLv3]
+
| UNLICENSE
 
| ESRGAN
 
| ESRGAN
 
| Anime/Pretrained
 
| Anime/Pretrained
 
| 2020-04-08
 
| 2020-04-08
| Anime landscape upscale. Trained on Frames from BluRay of Byousoku 5 Centimeter
+
| NTSC DVD-spec encode x4 scale super-resolution for Anime Drawing style content. The dataset has a LOT of data throughout almost every frame, so it had a lot of stuff to learn. The resulting DVD-spec encode also had some blocking at times so it also learned to fight off blocking.
| [https://mega.nz/#!vtgSWKQT!K7Asn2zKe4N70R2aV89KEMTKhH3aiyGAAiuQDJF09qs 2xESRGAN]
+
| RRDB_PSNR_x4
 
|  
 
|  
 
|-
 
|-
  
| [https://1fichier.com/?tsmmdifzgwkwtdfcc5uo 4x_OLDIES_290000_G_FINAL]
 
| solidd93110
 
| 4x
 
| WTFPL
 
| ESRGAN
 
| Anime
 
| 2020-08-16
 
| i made this model to upscale old anime and denoise.
 
| RRDB_PSNR_x4.pth
 
 
|
 
 
|-
 
|-
| [https://www23.zippyshare.com/v/lhOStpVa/file.html fidelbd_pokemodel]
+
| <small>[https://mega.nz/file/taBFxAAR#6wW-DrGh6oy2uz6aAKJpa36ngE2eKcDNA8Psy11oK10 2x_AnimeClassics_UltraLite_510K]</small>
| Neo-Raws#4055
+
| CG1989
 
| 2x
 
| 2x
| WTFPL
+
| CC BY-NC-SA 4.0
 
| ESRGAN
 
| ESRGAN
| Anime
+
| Anime/Pretrained
| 2020-08-29
+
| 2022-03-08
| Made this model to upscale old anime that looks blurry.
+
| A 2x Ultra Lite model coming in under 8MB. Trained with over 15 sets of LRs ranging in a wide amount of issues. Handles Rainbows, Dot Crawl, MPEG/H.264 Compression, and may even assist in removing halos, and fixing blurriness in certain cases. This is my first public model for everyone. Best when used on old anime that is grainy. I can't say what anime it's best suited for as I have tried multiple series, and have found it does a good job on most all the tests. I wouldn't say use this for Western Animation, but it may work. I have done a few tests that I have shown in the upscale results, but that was chained with other models to achieve such a result. This model is meant to retain the more natural look of a series. There is a color shift on the end result, not drastic, but still noticable. I figure you should fix any color issues in post that way to give a more polished upscale. Big thanks to @SaurusX for the model name, and just helping out in general with anything.
| 2xESRGAN
+
| 2X_DigitalFilmV5_Lite.pth
 +
| [https://imgsli.com/OTg3MjM Sample 1]
 +
 
 +
[https://imgsli.com/OTg3MjQ Sample 2]
 +
 
 +
[https://imgsli.com/OTg3MjU Sample 3]
 +
 
 +
[https://imgsli.com/OTg3Nzg Sample 4]
  
| https://imgsli.com/MjExMjQ/
 
 
|-
 
|-
  
| [https://1fichier.com/?u7kwdxdn6uljha3icte8 4x_OLDIES_ALTERNATIVE_FINAL.pth]
+
|-
| solidd93110
+
| [https://drive.google.com/file/d/1DRyqqy24OU6G_mwsGjy6Zaxan_Fy1LNU/view?usp=sharing 600k] [https://drive.google.com/file/d/1psKp2DLWscPmDuzSTC8XqXL0LWnWuTYs/view?usp=sharing 650k] BooruGan
 +
| Tal
 
| 4x
 
| 4x
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
+
| WTFPL
 
| ESRGAN
 
| ESRGAN
 
| Anime
 
| Anime
| 2020-09-01
+
| 2021-11-26
| This model was made for my project captain tsubasa anime so i don't know if it works good for anything else. just try it ;)
+
| This model is designed to mainly upscale anime artworks. If you have issues with chroma then try the 600k iterations release.
| RRDB_PSNR_x4.pth
+
| 4x_Manga109Attempt
 +
| [https://cdn.discordapp.com/attachments/579685650824036387/913500929012150323/4f012e36d686033b274261295da89b32-comparison.jpg Sample 1]
 +
 
 +
[https://cdn.discordapp.com/attachments/579685650824036387/913500929356070912/256x256fdbb-comparison.jpg Sample 2]
  
| https://imgsli.com/MjEzMDU
+
[https://cdn.discordapp.com/attachments/579685650824036387/913500929649688617/franky-super-810x456-comparison.jpg Sample 3]
 
|-
 
|-
  
| [https://1fichier.com/?cnw3fwws08tdoaxycimj 2x_SHARP_ANIME_V1]
+
|-
| solidd93110
+
| [https://drive.google.com/file/d/1FWzEgqaSg2z0qOSjcmAxEUB3clEkm1PG/view 2x_Byousoku_5_Centimeter.pth]
 +
| Mystery_Bullet#0642 (Shady Adel)
 
| 2x
 
| 2x
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
+
| [https://www.gnu.org/licenses/gpl-3.0.en.html GNU GPLv3]
 
| ESRGAN
 
| ESRGAN
| Anime
+
| Anime/Pretrained
| 2020-09-12
+
| 2020-04-08
| this model has been trained to work on lines and details - works well on animes which have fairly fine lines at the base but also the video must be progressive or deinterlaced
+
| Anime landscape upscale. Trained on Frames from BluRay of Byousoku 5 Centimeter
| 2xPSNR.pth
+
| [https://mega.nz/#!vtgSWKQT!K7Asn2zKe4N70R2aV89KEMTKhH3aiyGAAiuQDJF09qs 2xESRGAN]
 +
|
 +
|-
  
| https://imgsli.com/MjIwMDY
+
| [https://1fichier.com/?tsmmdifzgwkwtdfcc5uo 4x_OLDIES_290000_G_FINAL]
|-
+
| solidd93110
 
 
| [https://s.katou.pw/4x_MeguUp_150000_G.pth MeguUp]
 
| katoumegumi_#3231
 
|
 
| [https://www.gnu.org/licenses/gpl-3.0.en.html GNU GPLv3].
 
| ESRGAN
 
| Anime
 
| 2020-09-16
 
| Upscaling of lossless (uncompressed) anime art.
 
| Interpolated custom, hence the license.
 
|  SYNLA: LyonHrt (bloc97 dataset, MIT license).
 
Fatal Pixels, Fatality Anime: twittman (CC BY-NC-SA 4.0, acquired special permission to relicense. https://s.katou.pw/Discord_dPUiI1x3Pf.png).
 
Deviance: BlackScout (GNU GPLv3)
 
|-
 
 
 
 
 
| [https://icedrive.net/1/43GNBihZyi NMKD UltraYandere]
 
| |[[User:nmkd|Nmkd]]
 
 
| 4x
 
| 4x
 
| WTFPL
 
| WTFPL
 
| ESRGAN
 
| ESRGAN
| Art/Anime
 
| 2020-10-08
 
| Highly flexible 2D Art upscaling
 
| 4xESRGAN
 
 
|
 
|-
 
 
| [https://icedrive.net/1/43GNBihZyi NMKD UltraYandere Lite]
 
| |[[User:nmkd|Nmkd]]
 
| 4x
 
| WTFPL
 
| ESRGAN Lite [nf=32 nb=12]
 
 
| Anime
 
| Anime
| 2020-10-15
+
| 2020-08-16
| Fast Anime/Art upscaling
+
| i made this model to upscale old anime and denoise.
| 4x_DIV2K-Lite
+
| RRDB_PSNR_x4.pth
  
 
|  
 
|  
 
|-
 
|-
 
+
| [https://www23.zippyshare.com/v/lhOStpVa/file.html fidelbd_pokemodel]
| [https://www115.zippyshare.com/v/O2do1VCf/file.html 2x_pokemodel_lite_100000_G]
 
 
| Neo-Raws#4055
 
| Neo-Raws#4055
 
| 2x
 
| 2x
Line 913: Line 1,008:
 
| ESRGAN
 
| ESRGAN
 
| Anime
 
| Anime
| 2020-11-01
+
| 2020-08-29
| Upscale old anime like pokemon
+
| Made this model to upscale old anime that looks blurry.
| none
+
| 2xESRGAN
  
| https://imgsli.com/Mjc3Nzk https://imgsli.com/Mjc3Nzg https://imgsli.com/Mjc3ODA https://imgsli.com/Mjc3ODE https://imgsli.com/Mjc3ODI
+
| https://imgsli.com/MjExMjQ/
 
|-
 
|-
  
 
+
| <small>[https://1fichier.com/?u7kwdxdn6uljha3icte8 4x_OLDIES_ALTERNATIVE_FINAL.pth]</small>
| [https://1fichier.com/?020t1y83kjrusd95dh0a 2x_SHARP_ANIME_V2]
 
 
| solidd93110
 
| solidd93110
| 2x
+
| 4x
 
|[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]
 
| ESRGAN
 
| ESRGAN
 
| Anime
 
| Anime
| 2020-12-09
+
| 2020-09-01
| 2x_PSNR.pth
+
| This model was made for my project captain tsubasa anime so i don't know if it works good for anything else. just try it ;)
|
+
| RRDB_PSNR_x4.pth
| https://imgsli.com/MzI0MDk
+
 
 +
| https://imgsli.com/MjEzMDU
 
|-
 
|-
  
| [https://1fichier.com/?23hzkj20isef5b75cw3y 2x_BIGOLDIES_415000_G.pth]
+
| [https://1fichier.com/?cnw3fwws08tdoaxycimj 2x_SHARP_ANIME_V1]
 
| solidd93110
 
| solidd93110
 
| 2x
 
| 2x
Line 939: Line 1,034:
 
| ESRGAN
 
| ESRGAN
 
| Anime
 
| Anime
| 2020-12-09
+
| 2020-09-12
| upscaling old anime. help to denoise and find lines and dehalo
+
| this model has been trained to work on lines and details - works well on animes which have fairly fine lines at the base but also the video must be progressive or deinterlaced
| 2x_PSNR.pth
+
| 2xPSNR.pth
| https://imgsli.com/MzI0MDc
+
 
 +
| https://imgsli.com/MjIwMDY
 
|-
 
|-
  
 
+
| [https://s.katou.pw/4x_MeguUp_150000_G.pth MeguUp]
| [https://icedrive.net/1/f0UAiRqz3N NMKD YandereNeo]
+
| katoumegumi_#3231
| |[[User:nmkd|Nmkd]]
+
|
| 2x
+
| [https://www.gnu.org/licenses/gpl-3.0.en.html GNU GPLv3].
| WTFPL
+
| ESRGAN
| ESRGAN (Lite)
 
 
| Anime
 
| Anime
| 2021-01-26
+
| 2020-09-16
|  
+
| Upscaling of lossless (uncompressed) anime art.
|  
+
| Interpolated custom, hence the license.
| https://i.imgur.com/oxs71v5.png
+
| SYNLA: LyonHrt (bloc97 dataset, MIT license).
 +
Fatal Pixels, Fatality Anime: twittman (CC BY-NC-SA 4.0, acquired special permission to relicense. [https://s.katou.pw/Discord_dPUiI1x3Pf.png Screencap]).
 +
Deviance: BlackScout (GNU GPLv3)
 +
|-
 +
 
  
 +
| [https://icedrive.net/1/43GNBihZyi NMKD UltraYandere]
 +
| |[[User:nmkd|Nmkd]]
 +
| 4x
 +
| WTFPL
 +
| ESRGAN
 +
| Art/Anime
 +
| 2020-10-08
 +
| Highly flexible 2D Art upscaling
 +
| 4xESRGAN
  
 +
|
 
|-
 
|-
| [https://icedrive.net/1/f0UAiRqz3N NMKD YandereNeo]
+
 
 +
| [https://icedrive.net/1/43GNBihZyi NMKD UltraYandere Lite]
 
| |[[User:nmkd|Nmkd]]
 
| |[[User:nmkd|Nmkd]]
 
| 4x
 
| 4x
 
| WTFPL
 
| WTFPL
| ESRGAN (Lite)
+
| ESRGAN Lite [nf=32 nb=12]
 
| Anime
 
| Anime
| 2021-01-26
+
| 2020-10-15
 +
| Fast Anime/Art upscaling
 +
| 4x_DIV2K-Lite
 +
 
 
|  
 
|  
| 4x_DIV2K-Lite_1M
 
| https://i.imgur.com/oxs71v5.png
 
 
|-
 
|-
  
| [https://u.pcloud.link/publink/show?code=XZRAmuXZRy3vkvvBdTftheNOFeJ0tLLv74eX 2x_Waifaux-NL3-SuperLite]
+
| [https://www115.zippyshare.com/v/O2do1VCf/file.html 2x_pokemodel_lite_100000_G]
| Joey
+
| Neo-Raws#4055
 
| 2x
 
| 2x
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
+
| WTFPL
 
| ESRGAN
 
| ESRGAN
|  
+
| Anime
| 2021-02-23
+
| 2020-11-01
| Trained this model to see how it would work trying to essentially get the same results as Waifu2x but with ESRGAN
+
| Upscale old anime like pokemon
 
| none
 
| none
|
 
|-
 
  
| [https://u.pcloud.link/publink/show?code=XZPLWuXZbQz9nPrUhlytIN0ne4UjyzQRMHmX 2x_Waifaux-NL3-SRResNet]
+
| https://imgsli.com/Mjc3Nzk
| Joey
+
 
| 2x
+
https://imgsli.com/Mjc3Nzg
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
+
 
| EDSR (SRResNet)
+
https://imgsli.com/Mjc3ODA
|
+
 
| 2021-02-25
+
https://imgsli.com/Mjc3ODE
| Emulating Waifu2x at Noise Level 3 NOTE: You can't use this with regular esrgan forks or the bot, it has to be run through basicsr
+
 
| none
+
https://imgsli.com/Mjc3ODI
|
 
 
|-
 
|-
  
| [https://u.pcloud.link/publink/show?code=kZcUlBXZ9egcFjtf3cjX02lDqNAr5j4l3jFX 4x_Training4Melozard_Anime]
 
| Joey
 
| 4x
 
| Whatever @B.Melozard2 wants
 
| ESRGAN
 
| Anime
 
| 2021-03-17
 
|
 
| RRDB_ESRGAN_x4_old_arch
 
|
 
|-
 
  
| [https://drive.google.com/file/d/18iWj4eWiMfUd2WyLCEzMlZK1JVT7L8nT/view?usp=sharing 2x_LD-Anime_Skr_v1.0]
+
| [https://1fichier.com/?020t1y83kjrusd95dh0a 2x_SHARP_ANIME_V2]
| Skr
+
| solidd93110
 
| 2x
 
| 2x
 
|[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]
 
| ESRGAN
 
| ESRGAN
| Denoise/Dehalo
+
| Anime
| 2021-04-17
+
| 2020-12-09
| Denoise, dehalo, derainbow old anime
+
| 2x_PSNR.pth
| 2xESRGAN
+
|
| https://imgsli.com/NTA3MDU/
+
| https://imgsli.com/MzI0MDk
 
|-
 
|-
  
| [https://drive.google.com/drive/u/4/folders/19IQNM5Vo5vGcoKaRJaRrLScfxBbFBVff 2x_KemonoScale_v2]
+
| [https://1fichier.com/?23hzkj20isef5b75cw3y 2x_BIGOLDIES_415000_G.pth]
| EzoGaming
+
| solidd93110
 
| 2x
 
| 2x
 
|[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]
 
| ESRGAN
 
| ESRGAN
 
| Anime
 
| Anime
| 2021-06-05
+
| 2020-12-09
| Upscaling frames from Irodori anime (namely kemono friends) from 540p (the source render resolution) to 1080p, low resolution flat shaded art, de-JPEG of the aforementioned
+
| upscaling old anime. help to denoise and find lines and dehalo
| x2_CGIMaster_v1
+
| 2x_PSNR.pth
| https://cdn.discordapp.com/attachments/705815801290293329/850614850735439892/unknown.png
+
| https://imgsli.com/MzI0MDc
 
|-
 
|-
  
| [https://s.katou.pw/4x_muy4_035_1.pth 4x_muy4_035_1.pth]
+
 
| katoumegumi_
+
| [https://icedrive.net/1/f0UAiRqz3N NMKD YandereNeo]
 +
| |[[User:nmkd|Nmkd]]
 +
| 2x
 +
| WTFPL
 +
| ESRGAN (Lite)
 +
| Anime
 +
| 2021-01-26
 +
|
 +
|
 +
| https://i.imgur.com/oxs71v5.png
 +
 
 +
 
 +
|-
 +
| [https://icedrive.net/1/f0UAiRqz3N NMKD YandereNeo]
 +
| |[[User:nmkd|Nmkd]]
 
| 4x
 
| 4x
 
| WTFPL
 
| WTFPL
| ESRGAN
+
| ESRGAN (Lite)
 
| Anime
 
| Anime
| 2021-07-18
+
| 2021-01-26
| Upscaling of anime art (Specifically visual novel CG art)
 
 
|  
 
|  
| https://cdn.knightlab.com/libs/juxtapose/latest/embed/index.html?uid=278999ec-e802-11eb-abb7-b9a7ff2ee17c (Hakurei Reimu from Touhou Project illustrated by "okawa friend") https://cdn.knightlab.com/libs/juxtapose/latest/embed/index.html?uid=f3d4781a-e801-11eb-abb7-b9a7ff2ee17c (Kitaooji Karen from Making * Lovers developed by SMEE.)(edited)
+
| 4x_DIV2K-Lite_1M
 +
| https://i.imgur.com/oxs71v5.png
 
|-
 
|-
  
| [https://drive.google.com/file/d/1bqyG9llxkJ6i6MJTaNUiSgSXtrlPi0d7/view Falcon Fanart]
+
| [https://u.pcloud.link/publink/show?code=XZRAmuXZRy3vkvvBdTftheNOFeJ0tLLv74eX 2x_Waifaux-NL3-SuperLite]
| |[[User:LyonHrt|LyonHrt]]
+
| Joey
| 4x
+
| 2x
 
|[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]
 
| ESRGAN
 
| ESRGAN
| Anime
 
| not on discord
 
 
|  
 
|  
| 4xPSNR
+
| 2021-02-23
 +
| Trained this model to see how it would work trying to essentially get the same results as Waifu2x but with ESRGAN
 +
| none
 
|  
 
|  
 
|-
 
|-
|}
 
  
====Manga====
+
| [https://u.pcloud.link/publink/show?code=XZPLWuXZbQz9nPrUhlytIN0ne4UjyzQRMHmX 2x_Waifaux-NL3-SRResNet]
{| class="wikitable sortable"
+
| Joey
 +
| 2x
 +
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 +
| EDSR (SRResNet)
 +
|
 +
| 2021-02-25
 +
| Emulating Waifu2x at Noise Level 3 NOTE: You can't use this with regular esrgan forks or the bot, it has to be run through basicsr
 +
| none
 +
|  
 
|-
 
|-
! Model Name
 
! Author
 
! Scale
 
! License
 
! Architecture
 
! Purpose (short)
 
! Date Posted
 
! Purpose (Full)
 
! Pretrained_Model_G
 
! sample
 
  
|-
+
| [https://u.pcloud.link/publink/show?code=kZcUlBXZ9egcFjtf3cjX02lDqNAr5j4l3jFX 4x_Training4Melozard_Anime]
| [https://drive.google.com/file/d/1uraeNuUPx46HKZ77xUI1PCBBoH-LhikY/view?usp=sharing FSMangaV2]
+
| Joey
| Jacob_
 
 
| 4x
 
| 4x
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
+
| Whatever @B.Melozard2 wants
 
| ESRGAN
 
| ESRGAN
| Manga
+
| Anime
| 2020-12-14
+
| 2021-03-17
| Manga-style images with or without dithering - cartoons, maybe pixel art, etc
+
|  
| RRDB_PSNR_x4
+
| RRDB_ESRGAN_x4_old_arch
| https://i.imgur.com/oVBVthF.png
+
|  
 
|-
 
|-
  
|-
+
| [https://drive.google.com/file/d/18iWj4eWiMfUd2WyLCEzMlZK1JVT7L8nT/view?usp=sharing 2x_LD-Anime_Skr_v1.0]
| [https://cdn.discordapp.com/attachments/450036622940045325/936099965300768808/4x_eula_digimanga_bw_v1_860k.pth 4x_eula_digimanga_bw_v1_860k]
+
| Skr
| end user license agreement#9756
+
| 2x
| 4x
 
 
|[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]
 
| ESRGAN
 
| ESRGAN
| Manga
+
| Denoise/Dehalo
| 2022-1-26
+
| 2021-04-17
| Black and white digital manga with halftones.
+
| Denoise, dehalo, derainbow old anime
| 4xPSNR
+
| 2xESRGAN
| [https://slow.pics/c/lp9O3GnZ Clean] [https://slow.pics/c/AxmHpF3m Heavy Degradation]
+
| https://imgsli.com/NTA3MDU/
 
|-
 
|-
  
 +
| [https://drive.google.com/drive/u/4/folders/19IQNM5Vo5vGcoKaRJaRrLScfxBbFBVff 2x_KemonoScale_v2]
 +
| EzoGaming
 +
| 2x
 +
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 +
| ESRGAN
 +
| Anime
 +
| 2021-06-05
 +
| Upscaling frames from Irodori anime (namely kemono friends) from 540p (the source render resolution) to 1080p, low resolution flat shaded art, de-JPEG of the aforementioned
 +
| x2_CGIMaster_v1
 +
| [https://cdn.discordapp.com/attachments/705815801290293329/850614850735439892/unknown.png Sample]
 
|-
 
|-
| [https://cdn.discordapp.com/attachments/894310505622155264/1008498310597857290/4x_eula_digimanga_bw_v2_nc1_307k.pth 4x_eula_digimanga_bw_v2_nc1_307k]
+
 
| end user license agreement#9756
+
| [https://s.katou.pw/4x_muy4_035_1.pth 4x_muy4_035_1.pth]
 +
| katoumegumi_
 
| 4x
 
| 4x
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
+
| WTFPL
 
| ESRGAN
 
| ESRGAN
| Manga
+
| Anime
| 2022-08-17
+
| 2021-07-18
|Vast improvement over v1 in low frequency detail; moiré and artifacting reduced significantly and less random noise from JPEG artefacts in the input. Also now only works on 1 channel images, so it runs slightly faster on average and resulting images are much smaller but it might not work on some ESRGAN implementations, I personally recommend using chaiNNer. v1 may still be better in some edge cases.
+
| Upscaling of anime art (Specifically visual novel CG art)
There's also a supplementary 1x model that denoises very low quality LRs and smooths halftones so the image works better with the 4x model. Only trained it to help build the dataset and it's useless for already decent-ish LRs but may help you in some situations.
+
|
| 4x_eula_digimanga_bw_v1_860k
+
| [https://cdn.knightlab.com/libs/juxtapose/latest/embed/index.html?uid=278999ec-e802-11eb-abb7-b9a7ff2ee17c Sample 1]
| https://slow.pics/c/Asbu0xgz
+
 
 +
(Hakurei Reimu from Touhou Project illustrated by "okawa friend")
 +
 
 +
[https://cdn.knightlab.com/libs/juxtapose/latest/embed/index.html?uid=f3d4781a-e801-11eb-abb7-b9a7ff2ee17c Sample 2]
 +
 
 +
(Kitaooji Karen from Making * Lovers developed by SMEE.)(edited)
 
|-
 
|-
  
| [https://mega.nz/file/nR51EYYC#RcAKkxb8hAzvZAj4TLxnAukd9gbrYuUeyjUp2l1ffb8 MangaScaleV3]
+
| [https://drive.google.com/file/d/1bqyG9llxkJ6i6MJTaNUiSgSXtrlPi0d7/view Falcon Fanart]
| Bunzero++
+
| |[[User:LyonHrt|LyonHrt]]
| 2x
+
| 4x
 
|[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]
| ESRGAN+
 
| Manga
 
| 2021-05-24
 
| To upscale manga including halftones, instead of trying to smooth them out.
 
| unreleased model
 
| https://imgsli.com/NTUyMjg
 
|-
 
 
| [http://www.mediafire.com/file/w3jujtm752hvdj1/Manga109Attempt.pth.zip/file Manga109Attempt]
 
| Kingdomakrillic
 
| 4x
 
| [https://creativecommons.org/licenses/by/4.0/ CC BY 4.0]
 
 
| ESRGAN
 
| ESRGAN
| Anime / Manga
+
| Anime
|
+
| not on discord
| Pretrained model: [https://drive.google.com/drive/folders/1ldwajXL50uC7PCS63B4Wato6Dnk-svNL 4xPSNR]
+
|  
 +
| 4xPSNR
 
|  
 
|  
| Manga109
 
|-
 
|}
 
 
====Cartoons====
 
{| class="wikitable sortable"
 
|-
 
! Model Name
 
! Author
 
! Scale
 
! License
 
! Architecture
 
! Purpose (short)
 
! Date Posted
 
! Purpose (Full)
 
! Pretrained_Model_G
 
! sample
 
 
|-
 
|-
  
Line 1,161: Line 1,255:
 
| Improvement of low-quality cartoons from broadcast sources.
 
| Improvement of low-quality cartoons from broadcast sources.
 
| Will greatly increase the visual quality of bad broadcast tape sources of '80s and '90s cartoons (e.g. Garfield and Friends, Heathcliff, DuckTales, etc). Directly addresses chroma blur, dot crawl, and rainbowing. You're highly advised to take care of haloing beforehand in your favorite video editor as the model will not fix it and may make existing halos more noticeable.
 
| Will greatly increase the visual quality of bad broadcast tape sources of '80s and '90s cartoons (e.g. Garfield and Friends, Heathcliff, DuckTales, etc). Directly addresses chroma blur, dot crawl, and rainbowing. You're highly advised to take care of haloing beforehand in your favorite video editor as the model will not fix it and may make existing halos more noticeable.
| [https://imgsli.com/OTg0MjA/6/7 Sample 1] [https://imgsli.com/OTc1OTg/2/3 Sample 2] [https://imgsli.com/OTg0Mjc/0/1 Sample 3]
+
| [https://imgsli.com/OTg0MjA/6/7 Sample 1]
 +
 
 +
[https://imgsli.com/OTc1OTg/2/3 Sample 2]
 +
 
 +
[https://imgsli.com/OTg0Mjc/0/1 Sample 3]
 
|-
 
|-
  
Line 1,173: Line 1,271:
 
| In addition to removing the vertical blur, the model upscales, sharpens and will remove MPEG-2 artifacting and a small amount of rainbowing and dot crawl. Another series afflicted with the vertical blur is Avatar the Last Airbender, which can be repaired by this model. The video fed into the model MUST be 540 vertical for the deblur to work properly.
 
| In addition to removing the vertical blur, the model upscales, sharpens and will remove MPEG-2 artifacting and a small amount of rainbowing and dot crawl. Another series afflicted with the vertical blur is Avatar the Last Airbender, which can be repaired by this model. The video fed into the model MUST be 540 vertical for the deblur to work properly.
 
| 2xESRGAN
 
| 2xESRGAN
| https://imgsli.com/NzM2MDY https://imgsli.com/ODIxMjU
+
| https://imgsli.com/NzM2MDY  
 +
 
 +
https://imgsli.com/ODIxMjU
 
|-
 
|-
  
Line 1,199: Line 1,299:
 
| This model was trained to restore "Sheeep" while retaining and enhancing the noise present in the show. The model amazingly finished training in only 8.8k iterations with no pretrain, with a dataset of only 4 image pairs. This model should work well for animes and cartoons with a lot of grain present. There are some slight haloing issues in dark colors unfortunately, but I was unable to fix it.
 
| This model was trained to restore "Sheeep" while retaining and enhancing the noise present in the show. The model amazingly finished training in only 8.8k iterations with no pretrain, with a dataset of only 4 image pairs. This model should work well for animes and cartoons with a lot of grain present. There are some slight haloing issues in dark colors unfortunately, but I was unable to fix it.
 
| [https://cdn.discordapp.com/attachments/903415274521374750/999820017652740176/1658446060.8526323.png Sample 1]
 
| [https://cdn.discordapp.com/attachments/903415274521374750/999820017652740176/1658446060.8526323.png Sample 1]
 +
 
[https://cdn.discordapp.com/attachments/903415274521374750/999821036583391292/1658446301.0955725.png Sample 2]
 
[https://cdn.discordapp.com/attachments/903415274521374750/999821036583391292/1658446301.0955725.png Sample 2]
 +
 
[https://cdn.discordapp.com/attachments/903415274521374750/999821212718993478/1658446346.8536465.png Sample 3]
 
[https://cdn.discordapp.com/attachments/903415274521374750/999821212718993478/1658446346.8536465.png Sample 3]
 +
 
[https://cdn.discordapp.com/attachments/903415274521374750/999820756043169803/1658446237.0734627.png Sample 4]
 
[https://cdn.discordapp.com/attachments/903415274521374750/999820756043169803/1658446237.0734627.png Sample 4]
 
|-
 
|-
  
 
|-
 
|-
| [https://mega.nz/folder/eBwyyKbA#CApT1_CrbpGBymLmc-FCTQ 2x-UniScale_CartoonRestore-lite]
+
| <small>[https://mega.nz/folder/eBwyyKbA#CApT1_CrbpGBymLmc-FCTQ 2x-UniScale_CartoonRestore-lite]</small>
 
| [[User:Kim2091|Kim2091]]
 
| [[User:Kim2091|Kim2091]]
 
| 2x
 
| 2x
Line 1,216: Line 1,319:
 
| Comparisons:
 
| Comparisons:
 
[https://cdn.discordapp.com/attachments/880826637543964712/882136618461458462/unknown.png Sample 1]
 
[https://cdn.discordapp.com/attachments/880826637543964712/882136618461458462/unknown.png Sample 1]
 +
 
[https://cdn.discordapp.com/attachments/880826637543964712/882136455999262750/unknown.png Sample 2]
 
[https://cdn.discordapp.com/attachments/880826637543964712/882136455999262750/unknown.png Sample 2]
 +
 
[https://cdn.discordapp.com/attachments/880826637543964712/882136979410673684/unknown.png Sample 3]
 
[https://cdn.discordapp.com/attachments/880826637543964712/882136979410673684/unknown.png Sample 3]
 +
 
[https://cdn.discordapp.com/attachments/649861104645701637/882121519973666886/unknown.png Sample 4]
 
[https://cdn.discordapp.com/attachments/649861104645701637/882121519973666886/unknown.png Sample 4]
 +
 
[https://cdn.discordapp.com/attachments/649861104645701637/882124248368427078/unknown.png Sample 5]
 
[https://cdn.discordapp.com/attachments/649861104645701637/882124248368427078/unknown.png Sample 5]
 +
 
[https://cdn.discordapp.com/attachments/880826637543964712/882110147676221480/Spongebob_UniScale_CartoonRestore.mp4 Video Sample]
 
[https://cdn.discordapp.com/attachments/880826637543964712/882110147676221480/Spongebob_UniScale_CartoonRestore.mp4 Video Sample]
 +
 
[https://cdn.discordapp.com/attachments/547949806761410560/881971445813633044/Spongebob_Original.mp4 Original Video]
 
[https://cdn.discordapp.com/attachments/547949806761410560/881971445813633044/Spongebob_Original.mp4 Original Video]
 
|-
 
|-
  
| [https://www108.zippyshare.com/v/PYWFtETd/file.html 2x_ATLA_KORRA_336200_G.pth]
+
| <small>[https://www108.zippyshare.com/v/PYWFtETd/file.html 2x_ATLA_KORRA_336200_G.pth]</small>
 
| aptitude
 
| aptitude
 
| 2x
 
| 2x
Line 1,233: Line 1,342:
 
| Upscaling of Animation based on The Legend of Korra.
 
| Upscaling of Animation based on The Legend of Korra.
 
| 2xESRGAN.pth
 
| 2xESRGAN.pth
| https://imgsli.com/NjIzMzU https://imgsli.com/NjIzMTY https://imgsli.com/NjIzMTc https://imgsli.com/NjIzMTg https://imgsli.com/NjIzMTk https://imgsli.com/NjIzMjA https://imgsli.com/NjIzMjE https://imgsli.com/NjIzMjQ https://imgsli.com/NjIzMjc https://imgsli.com/NjIzMjY https://imgsli.com/NjIzMjU
+
| https://imgsli.com/NjIzMzU  
 +
 
 +
https://imgsli.com/NjIzMTY  
 +
 
 +
https://imgsli.com/NjIzMTc  
 +
 
 +
https://imgsli.com/NjIzMTg  
 +
 
 +
https://imgsli.com/NjIzMTk  
 +
 
 +
https://imgsli.com/NjIzMjA  
 +
 
 +
https://imgsli.com/NjIzMjE  
 +
 
 +
https://imgsli.com/NjIzMjQ  
 +
 
 +
https://imgsli.com/NjIzMjc  
 +
 
 +
https://imgsli.com/NjIzMjY  
 +
 
 +
https://imgsli.com/NjIzMjU
 
|-
 
|-
  
Line 1,251: Line 1,380:
  
 
|-
 
|-
| [https://drive.google.com/file/d/1dT7Pw0iwPkQC4-rWriibrVLuA3KS3pS4/view SpongeBob (4x_SpongeBob_235000_G / 4xSpongebob)]
+
| <small>[https://drive.google.com/file/d/1dT7Pw0iwPkQC4-rWriibrVLuA3KS3pS4/view Spongebob (4x_SpongeBob_235000_G / 4xSpongebob)]</small>
 
| Joey
 
| Joey
 
| 4x
 
| 4x
Line 1,312: Line 1,441:
 
|-
 
|-
  
| [https://drive.google.com/open?id=1D3AjglmYlQ2HGda7iw4JXckSGbVBqsQr SpongeBob.CEL.2.HD.125ki.499e-PHOENiX.pth]
+
| <small>[https://drive.google.com/open?id=1D3AjglmYlQ2HGda7iw4JXckSGbVBqsQr SpongeBob.CEL.2.HD.125ki.499e-PHOENiX.pth]</small>
 
| Phoenix
 
| Phoenix
 
| 4x
 
| 4x
Line 1,359: Line 1,488:
 
|-
 
|-
  
|}
 
  
====Digital Animation====
 
{| class="wikitable sortable"
 
|-
 
! Model Name
 
! Author
 
! Scale
 
! License
 
! Architecture
 
! Purpose (short)
 
! Date Posted
 
! Purpose (Full)
 
! Pretrained_Model_G
 
! sample
 
  
 
|-
 
|-
| [https://u.pcloud.link/publink/show?code=XZQlqjXZPNgmLRJl5HLLG0M3VRmzmzmHr5hk 4x_DigitalFake-2.1]
+
| <small>[https://drive.google.com/file/d/1i68bWo7xh3JPFIO-zLv9UMD72E-5xNr_/view American Dad 2 (American.Dad.2.HD.150ki.5e-PHOENiX.pth)]</small>
| Joey
+
| PHEONiX (?)
 
| 4x
 
| 4x
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
+
|  
 
| ESRGAN
 
| ESRGAN
| Digital Animation
+
| Animation
| 2021-02-18
+
|  
| Replica of DigitalFrames 2.1 but interpolatable
+
| American Dad
| RRDB_ESRGAN_x4_old_arch.pth
+
|  
 
|  
 
|  
 
|-
 
|-
  
| [https://mega.nz/folder/QsUWWJjb#SIuqcPEzrkpLW0WNSth66g 4X_KCJPUNK_1.0233089 G.pth]
+
|-
| KCJPUNK
+
| [https://de-next.owncube.com/index.php/s/jYnFtncarkBmpcF FatalimiX]
 +
| |[[User:Twittman|Twittman]]
 
| 4x
 
| 4x
|[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] (?)
 
| ESRGAN
 
| ESRGAN
| Digital Animation
+
| Animation
| 2021-03-23
+
| 2019-10-29
| Up-scaling Digital Animation
+
| Comic and Cartoon style images
| 4x_fatal_Anime_500000_G
+
| 4x_Fatality_MKII_90000_G_02.pth
 
|  
 
|  
 
|-
 
|-
  
| [https://mega.nz/folder/4oVRHQaA#vE3Lplfetc9z9SJOxZD6nA 2X_KcjpunkAnime_2.0_Lite_196496_G]
+
|-
| KCJPUNK
+
| [https://drive.google.com/open?id=1qjpxp8z4FGLZxieUDEVKd6FDum-8vpvJ Comic Book]
| 2x
+
| |[[User:LyonHrt|LyonHrt]]
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
+
| 4x
| ESRGAN-lite
+
| [https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
| Digital Animation
+
| ESRGAN
| 2021-03-23
+
| Animation
| Up-scaling Digital Animation
+
|  
| This is my first attempt to make Light model so I started with 2x version. This model is much faster and give better results than my previous one.
+
| Comic / Drawings. Trained on Custom (Spider-Man) dataset 
 +
| none (no interpolation)
 
|  
 
|  
 
|-
 
|-
  
| [https://mega.nz/file/XWpFzRID#398qZx763bMSBsPB2MfgZqROGxdB0VldrIgH_prwa7U 2x_DigitoonLite_216k]
 
| SaurusX
 
| 2x
 
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 
| ESRGAN-lite
 
| Digital Animation
 
| 2022-07-05
 
| Meant as a versatile model for upscaling high detail digital anime and cartoons. Has debanding, MPEG-2 correction, and halo reduction. Trained to handle both 4:3 and 16:9 DVD material with equal efficacy. Will retain a lot of textures except for the really high freq stuff.
 
| 2x_DigiGradients_Lite_486k.pth
 
| https://imgsli.com/MTE1Mzg4
 
 
|-
 
|-
 
+
| [https://de-next.owncube.com/index.php/s/x99pKzS7TNaErrC Fatal_Anime]
| [https://mega.nz/file/PfhxWCRT#R59B_ovEwdiojZbas76SCbUSv-uNd1ZSOdZ0vQp-6w4 2x_DigiGradients_Lite_486k]
+
| |[[User:Twittman|Twittman]]
| SaurusX
+
| 4x
| 2x
+
| [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]
+
| ESRGAN
| ESRGAN-lite
+
| Animation
| Digital Animation
 
| 2022-08-24
 
| A very focused model meant for upscaling the TMNT 2003 DVDs. Degradations were added via AVISynth in order to match the video on the TMNT 2003 DVDs to correct the source problems. Problems corrected include aliased red chroma, chroma vertical blur, bad deinterlacing, banding, compression "grain", and poor animation line detail. The AVS scripts for the LR's were run through HCEnc to get authentic low bit rate MPEG-2 artifacts for fixing. By design, the final model gives a very digital-looking result and does not do a good job of retaining textures as the style of TMNT 2003 is all flats and gradients.
 
| none
 
| https://imgsli.com/MTAxNjc1 https://imgsli.com/ODQ2NzI
 
|-
 
|}
 
 
 
====Drawings====
 
 
 
{| class="wikitable sortable"
 
|-
 
! Model Name
 
! Author
 
! Scale
 
! License
 
! Architecture
 
! Purpose (short)
 
! Date Posted
 
! Purpose (Full)
 
! Pretrained_Model_G
 
! sample
 
 
 
 
 
|-
 
| [https://www.dropbox.com/s/52hsqa7ymt5pgga/KDM003_scans_1x.pth KDM003_scans_1x.pth]
 
| BoxDrop
 
| 1x
 
 
|  
 
|  
| ESRGAN
+
| Trained on Anime and Manga images
| Art
+
| none (no interpolation)
| 2019-07-29
 
| Clean up model for scanned illustrations. - Made to remove moire patterns, reduce small imperfections, and correct mild compression artifacts in scanned Kingdom Death: Monster illustrations. CMYK printing often shifts colors, so this is intended to reverse that color shifting as well.
 
| Failed attempts based on ESRGAN_1x_JPEG_80to100
 
 
|  
 
|  
 
|-
 
|-
  
 
|-
 
|-
| [https://drive.google.com/open?id=103MX2bvd3GW0MYpC53ABU5VYY6v1t99Y DigiPaint]
+
| [http://cloud.mitch.best/index.php/s/iio9SL9BdgWiCoi 4x_AbeScale]
| Rastrum
+
| |[[User:mitch#1337|mitch#1337]]
 
| 4x
 
| 4x
| CC0
+
| GNU
 
| ESRGAN
 
| ESRGAN
| Art
+
| Linework Cartoons
| 2019-09-07
+
|  
| Digital Art Upscaler
+
| Dataset: LR: Clone High DVD, HR: Re-illustrated Vectors from Frames
| 4xfalcoon300(manga).pth
+
| [https://mega.nz/#!uhAVEaAA!qfwIQ44Ba3rXMAb2-M3aM3zFBCwzmyQ64IO_0O-csJE 4xESRGAN]
 
|  
 
|  
 
|-
 
|-
|}
 
  
====General Animation====
+
|-
 +
| <small>[https://mega.nz/file/9dJFjYrT#LY4Op8myZGJsh6ytUvF6ITVGXpx8XODN_AG-kzvWg4Y 2x_Loyaldk-SuperPony_370000 V1.0]</small>
 +
| ChaseMMD
 +
| 2x
 +
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 +
| ESRGAN
 +
| Animation
 +
| 2021-02-24
 +
| Upscale MLP episodes
 +
| 2x_ERGAN
 +
| https://imgsli.com/NDIwNTY
 +
 
 +
https://imgsli.com/NDIwNTg
 +
 
 +
https://imgsli.com/NDIwNjE
 +
 
 +
https://imgsli.com/NDIwNjI
  
{| class="wikitable sortable"
+
https://imgsli.com/NDIwNjM
 
|-
 
|-
! Model Name
 
! Author
 
! Scale
 
! License
 
! Architecture
 
! Purpose (short)
 
! Date Posted
 
! Purpose (Full)
 
! Pretrained_Model_G
 
! sample
 
  
 
|-
 
|-
| [https://drive.google.com/file/d/1i68bWo7xh3JPFIO-zLv9UMD72E-5xNr_/view American Dad 2 (American.Dad.2.HD.150ki.5e-PHOENiX.pth)]
+
| <small>[https://mega.nz/file/RYxGEKRL#nTe0Zy2gB9e9QF5T8UhJSla1oEPQjSCXAB0BpNrTrRk 2x_Loyaldk-LitePony_380500 V1.0]</small>
| PHEONiX (?)
+
| ChaseMMD
| 4x
+
| 2x
|  
+
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 
| ESRGAN
 
| ESRGAN
 
| Animation
 
| Animation
 +
| 2021-03-01
 +
| Upscale MLP episodes. Liter version.
 +
| 2x_ESRGAN
 
|  
 
|  
| American Dad
+
|-
|
 
|
 
|-
 
  
 
|-
 
|-
| [https://de-next.owncube.com/index.php/s/jYnFtncarkBmpcF FatalimiX]
+
| <small>[https://mega.nz/file/VM5WmLwb#QvO1n6w6llYRh8qauXXkXy4tElrBCwclCl5VHM3IH8M 2x_Loyaldk-LitePony_500000_V2.0]</small>
| |[[User:Twittman|Twittman]]
+
| ChaseMMD#6957
| 4x
+
| 2x
| [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]
| ESRGAN
+
| ESRGAN (NF=32, NB=12)
 
| Animation
 
| Animation
| 2019-10-29
+
| 2021-04-06
| Comic and Cartoon style images
+
| Upscale MLP episodes. Liter version. Able to handle compression better compared to V1.0 and no longer creates halo's/rainbows. Good for Vector 2D art however converts detail to blobs.
| 4x_Fatality_MKII_90000_G_02.pth
+
| none
 
|  
 
|  
 
|-
 
|-
  
 
|-
 
|-
| [https://drive.google.com/open?id=1qjpxp8z4FGLZxieUDEVKd6FDum-8vpvJ Comic Book]
+
| <small>[https://mega.nz/file/YMwm0TbZ#agDBMDkswkpMM7FE1tDRlGX4zIOE4D0WBzcWudkN7Mc 2x_Loyaldk-MediumPony_500000_V2.0]</small>
| |[[User:LyonHrt|LyonHrt]]
+
| ChaseMMD
| 4x
+
| 2x
 
| [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]
| ESRGAN
+
| ESRGAN (NF=48, NB=18)
 
| Animation
 
| Animation
|  
+
| 2021-04-11
| Comic / Drawings. Trained on Custom (Spider-Man) dataset 
+
| Upscale MLP episodes. Liter version. Able to handle compression better compared to V1.0 and no longer creates halo's/rainbows. Good for Vector 2D art however converts detail to blobs. Leaves more blobs when converting detail to blobs compared to litepony
| none (no interpolation)
+
| none
 
|  
 
|  
 
|-
 
|-
  
 
|-
 
|-
| [https://de-next.owncube.com/index.php/s/x99pKzS7TNaErrC Fatal_Anime]
+
| <small>[https://mega.nz/file/hcoViIQQ#O1cmQ8FsLk5RAIP5lxB3Q2L0suh9IWX_dGHhuSuL9U4 4x_Loyaldk-LitePony_500000_V2.0]</small>
| |[[User:Twittman|Twittman]]
+
| ChaseMMD
 
| 4x
 
| 4x
| [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]
| ESRGAN
+
| ESRGAN (NF=32, NB=12)
 
| Animation
 
| Animation
|  
+
| 2021-04-12
| Trained on Anime and Manga images
+
| Good for Vector 2D art however converts detail to blobs. Quality seems weak with this one. However posting anyways to see if a use is found.
| none (no interpolation)
+
| none
 
|  
 
|  
 
|-
 
|-
  
 
|-
 
|-
| [https://1drv.ms/u/s!Aip-EMByJHY26wU-wOqApp7Hu7tt?e=4ihnBr HugePaint]
+
| <small>[https://mega.nz/file/IBhElRYJ#C_lVc8Vr2q8X6hEltJKD6I63U7oL_FDLeJgPh72AZmg 2x_Loyaldk-SuperPony_500000_V2.0]</small>
| |[[User:DinJerr|DinJerr]]
+
| ChaseMMD
| 8x
+
| 2x
| [https://creativecommons.org/licenses/by-nc/4.0/ CC BY-NC 4.0]
+
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 
| ESRGAN
 
| ESRGAN
| Digital Illustrations
+
| Animation
|  
+
| 2021-04-30
| Trained on a variety of images from ArtStation
+
| Upscale MLP episodes. Full version. Able to handle compression better compared to V1.0 and no longer creates halo's/rainbows. Good for Vector 2D art however converts detail to blobs.
| HugePeeps
+
| none
 
|  
 
|  
 
|-
 
|-
  
 
|-
 
|-
| [http://cloud.mitch.best/index.php/s/iio9SL9BdgWiCoi 4x_AbeScale]
+
| <small>[https://mega.nz/file/BMhUCbJZ#kYz0wvc26RW0n0f8AUXekW_ZCuRamKmh2-iWf1eefn0 4x_Loyaldk-SuperPony_500000_V2.0]</small>
| |[[User:mitch#1337|mitch#1337]]
+
| ChaseMMD
 
| 4x
 
| 4x
| GNU
+
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 
| ESRGAN
 
| ESRGAN
| Linework Cartoons
+
| Animation
|  
+
| 2021-04-30
| Dataset: LR: Clone High DVD, HR: Re-illustrated Vectors from Frames
+
| Upscale MLP episodes. Full version. Able to handle compression better compared to V1.0 and no longer creates halo's/rainbows. Good for Vector 2D art however converts detail to blobs.
| [https://mega.nz/#!uhAVEaAA!qfwIQ44Ba3rXMAb2-M3aM3zFBCwzmyQ64IO_0O-csJE 4xESRGAN]
+
| none
 
|  
 
|  
 
|-
 
|-
  
 
|-
 
|-
| [https://mega.nz/file/9dJFjYrT#LY4Op8myZGJsh6ytUvF6ITVGXpx8XODN_AG-kzvWg4Y 2x_Loyaldk-SuperPony_370000 V1.0]
+
| <small>[https://mega.nz/file/lYo01DRI#HIiJ7M5YaXjwAb3zEbtSLzAI59iljwEX72Ad88CvUf0 4x_Loyaldk-MediumPony_500000_V2.0]</small>
 
| ChaseMMD
 
| ChaseMMD
| 2x
+
| 4x
 
|[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]
 
| ESRGAN
 
| ESRGAN
 
| Animation
 
| Animation
| 2021-02-24
+
| 2021-04-30
| Upscale MLP episodes
+
| Upscale MLP episodes. Full version. Able to handle compression better compared to V1.0 and no longer creates halo's/rainbows. Good for Vector 2D art however converts detail to blobs.
| 2x_ERGAN
+
| none
| https://imgsli.com/NDIwNTY https://imgsli.com/NDIwNTg https://imgsli.com/NDIwNjE https://imgsli.com/NDIwNjI https://imgsli.com/NDIwNjM
+
|
 
|-
 
|-
  
 
|-
 
|-
| [https://mega.nz/file/RYxGEKRL#nTe0Zy2gB9e9QF5T8UhJSla1oEPQjSCXAB0BpNrTrRk 2x_Loyaldk-LitePony_380500 V1.0]
+
| [https://mega.nz/folder/5UAlQCRJ#2Bj1lq8PYeB3P9k5UKSvkQ Platoon Model Set]
 
| ChaseMMD
 
| ChaseMMD
| 2x
+
| 4x
 
|[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]
 
| ESRGAN
 
| ESRGAN
 
| Animation
 
| Animation
| 2021-03-01
+
| 2021-06-07
| Upscale MLP episodes. Liter version.
+
| Upscale Anime while keeping as much of the original detail as possible. Isn't among the sharpest models and if the anime is somewhat blurry it will retain that detail but add more pixels to give a smoother edge. Doesn't change the color very much and may make some scene's very slightly darker.
| 2x_ESRGAN
+
| Pony Models
 
|  
 
|  
 
|-
 
|-
  
 +
|}
 +
 +
=====Anime and Cartoon Restoration=====
 +
 +
{| class="wikitable sortable"
 
|-
 
|-
| [https://mega.nz/file/VM5WmLwb#QvO1n6w6llYRh8qauXXkXy4tElrBCwclCl5VHM3IH8M 2x_Loyaldk-LitePony_500000_V2.0]
+
! Model Name
| ChaseMMD#6957
+
! Author
| 2x
+
! Scale
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
+
! License
| ESRGAN (NF=32, NB=12)
+
! Architecture
| Animation
+
! Purpose (short)
| 2021-04-06
+
! Date Posted
| Upscale MLP episodes. Liter version. Able to handle compression better compared to V1.0 and no longer creates halo's/rainbows. Good for Vector 2D art however converts detail to blobs.
+
! Purpose (Full)
| none
+
! Pretrained_Model_G
|
+
! Sample
 
|-
 
|-
  
 
|-
 
|-
| [https://mega.nz/file/YMwm0TbZ#agDBMDkswkpMM7FE1tDRlGX4zIOE4D0WBzcWudkN7Mc 2x_Loyaldk-MediumPony_500000_V2.0]
+
| [https://mega.nz/file/MZdEja6I#ao8L6INGvzns4IlDpyzjpuzR-kIqm_VrC3Q8pUTZncM 1x_HurrDeblur_SuperUltraCompact]
| ChaseMMD
+
| Zarxrax
| 2x
+
| 1x
|  
+
|[http://www.wtfpl.net WTFPL]
| ESRGAN (NF=48, NB=18)
+
| Compact
 +
[nf:24 nc:8]
 
| Animation
 
| Animation
| 2021-04-11
+
| 2022-06-04
| Upscale MLP episodes. Liter version. Able to handle compression better compared to V1.0 and no longer creates halo's/rainbows. Good for Vector 2D art however converts detail to blobs. Leaves more blobs when converting detail to blobs compared to litepony
+
| This is a sharpening/deblurring model for anime video. It was created with three goals in mind:
| none
+
* Be blazing fast
 +
* Try to avoid enhancing noise and compression artifacts
 +
* Try to avoid over-enhancing intentionally blurred parts of the image
 +
Despite that last point, this is not intended to be used on modern anime which makes heavy of use depth-of-field effects.
 
|  
 
|  
|-
+
| https://imgsli.com/MTEwOTg4
  
 +
https://imgsli.com/MTEwOTY2
 +
 +
https://imgsli.com/MTEwOTY5
 +
 +
https://imgsli.com/MTEwOTcx
 
|-
 
|-
| [https://mega.nz/file/hcoViIQQ#O1cmQ8FsLk5RAIP5lxB3Q2L0suh9IWX_dGHhuSuL9U4 4x_Loyaldk-LitePony_500000_V2.0]
+
 
| ChaseMMD
+
| [https://mega.nz/file/9INFwYLD#KlV11g5yDG6fWWdZDXhztnsqhuLSIUP7-Upq3szgMCQ 1x_AnimeUndeint_Compact]
| 4x
+
| Zarxrax
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
+
| 1x
| ESRGAN (NF=32, NB=12)
+
|[http://www.wtfpl.net WTFPL]
 +
| Compact
 
| Animation
 
| Animation
| 2021-04-12
+
| 2022-06-05
| Good for Vector 2D art however converts detail to blobs. Quality seems weak with this one. However posting anyways to see if a use is found.
+
| This model corrects jagged lines on animation that has been deinterlaced. It handles simple line doubling, line interpolation, and even Yadif-style artifacts. It can also handle sources that were resized after deinterlacing, for example resizing from ntsc to pal resolutions. If a source has been upscaled after deinterlacing, it will need to be downsized before applying this model.
| none
 
 
|  
 
|  
 +
| https://imgsli.com/MTExMTE0
 +
 +
https://imgsli.com/MTExMTE1
 +
 +
https://imgsli.com/MTExMTE2
 +
 +
https://imgsli.com/MTExMTE3
 
|-
 
|-
  
|-
+
| [https://mega.nz/file/lNdTmSKD#R7GIISShx-zsUVwDHw0oosv72ifbB31GClXXV3UtLLo 1x_BleedOut_Compact]
| [https://mega.nz/file/IBhElRYJ#C_lVc8Vr2q8X6hEltJKD6I63U7oL_FDLeJgPh72AZmg 2x_Loyaldk-SuperPony_500000_V2.0]
+
| Zarxrax
| ChaseMMD
+
| 1x
| 2x
+
|[http://www.wtfpl.net WTFPL]
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
+
| Compact
| ESRGAN
 
 
| Animation
 
| Animation
| 2021-04-30
+
| 2022-07-25
| Upscale MLP episodes. Full version. Able to handle compression better compared to V1.0 and no longer creates halo's/rainbows. Good for Vector 2D art however converts detail to blobs.
+
| This model helps repair color bleed and heavy chroma noise that may be present on some older footage, particularly that which was recorded on VHS. It also cleans up rainbows if they are present.
| none
 
 
|  
 
|  
|-
+
| https://imgsli.com/MTE4MjEz
  
|-
+
https://imgsli.com/MTE4MjE4
| [https://mega.nz/file/BMhUCbJZ#kYz0wvc26RW0n0f8AUXekW_ZCuRamKmh2-iWf1eefn0 4x_Loyaldk-SuperPony_500000_V2.0]
 
| ChaseMMD
 
| 4x
 
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 
| ESRGAN
 
| Animation
 
| 2021-04-30
 
| Upscale MLP episodes. Full version. Able to handle compression better compared to V1.0 and no longer creates halo's/rainbows. Good for Vector 2D art however converts detail to blobs.
 
| none
 
|
 
|-
 
  
|-
+
https://imgsli.com/MTE4MjE5
| [https://mega.nz/file/lYo01DRI#HIiJ7M5YaXjwAb3zEbtSLzAI59iljwEX72Ad88CvUf0 4x_Loyaldk-MediumPony_500000_V2.0]
 
| ChaseMMD
 
| 4x
 
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 
| ESRGAN
 
| Animation
 
| 2021-04-30
 
| Upscale MLP episodes. Full version. Able to handle compression better compared to V1.0 and no longer creates halo's/rainbows. Good for Vector 2D art however converts detail to blobs.
 
| none
 
|
 
 
|-
 
|-
  
|-
+
| [https://mega.nz/file/VQkDmSgb#dlVuJZQ__yAKdJaWA_4Nx52YU_c2qy_a2oG9_a98dO4 1x_Dotzilla_Compact]
| [https://mega.nz/folder/5UAlQCRJ#2Bj1lq8PYeB3P9k5UKSvkQ Platoon Model Set]
+
| Zarxrax
| ChaseMMD
+
| 1x
| 4x
+
|[http://www.wtfpl.net WTFPL]
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
+
| Compact
| ESRGAN
 
 
| Animation
 
| Animation
| 2021-06-07
+
| 2022-12-09
| Upscale Anime while keeping as much of the original detail as possible. Isn't among the sharpest models and if the anime is somewhat blurry it will retain that detail but add more pixels to give a smoother edge. Doesn't change the color very much and may make some scene's very slightly darker.
+
| Wipes out dot crawl and rainbows in animation.
| Pony Models
+
| 1x_Compact_Pretrain.pth
| https://mega.nz/folder/5UAlQCRJ#2Bj1lq8PYeB3P9k5UKSvkQ
+
| https://imgsli.com/MTM4ODkz
 
|-
 
|-
 
|}
 
|}
  
 
+
====Digital Animation====
====Traditional Animation====
 
 
{| class="wikitable sortable"
 
{| class="wikitable sortable"
 
|-
 
|-
Line 1,704: Line 1,781:
 
! Purpose (Full)
 
! Purpose (Full)
 
! Pretrained_Model_G
 
! Pretrained_Model_G
! sample
+
! Sample
  
 
|-
 
|-
| [https://cdn.discordapp.com/attachments/579685650824036387/812414843939455036/2X_DigitalFilmV5_Lite.pth DigitalFilmV5 Lite]
+
| [https://u.pcloud.link/publink/show?code=XZQlqjXZPNgmLRJl5HLLG0M3VRmzmzmHr5hk 4x_DigitalFake-2.1]
| OptimusPrimal
+
| Joey
| 2x
+
| 4x
| WTFPL
+
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 
| ESRGAN
 
| ESRGAN
| Traditional Animation
+
| Digital Animation
| 2021-02-19
+
| 2021-02-18
| Upscaling Dragon Ball Z DBox DVD's. - grainy sources. It keeps some grain, but also doing some cleaning, sharpening, and fixing.
+
| Replica of DigitalFrames 2.1 but interpolatable
| none
+
| RRDB_ESRGAN_x4_old_arch.pth
| https://imgsli.com/NDE0MzQ https://imgsli.com/NDE0MzU https://imgsli.com/NDE0MzY https://imgsli.com/NDE0Mzc https://imgsli.com/NDE0Mzg https://imgsli.com/NDE0Mzk
+
|  
 
|-
 
|-
  
| [https://mega.nz/file/GFFkSL5R#C-dDxaxTZpSPUh_heAUw_idGIF8Xizdr0ktyGi05yFk 4x_HellinaCel]
+
| [https://mega.nz/folder/QsUWWJjb#SIuqcPEzrkpLW0WNSth66g 4X_KCJPUNK_1.0233089 G.pth]
| VGDCKeroro
+
| KCJPUNK
 
| 4x
 
| 4x
| fuckin do whatever
+
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 
| ESRGAN
 
| ESRGAN
| Traditional Animation
+
| Digital Animation
| 2021-03-21
+
| 2021-03-23
| A rougher alternative to 4xCelFrames with a focus on realistic looking cels over nice looking cels. It's trained on DetoriationFrames LRs, so if you give it an image straight from the source it will go twice as hard on it. This can be used to your advantage, though. I recommend cleaning it up before running DetoriationFrames, or else it will come out rough.
+
| Up-scaling Digital Animation
| 4xESRGAN
+
| 4x_fatal_Anime_500000_G
| https://imgsli.com/NDU2MjU
+
|  
 
|-
 
|-
|}
 
  
 +
| [https://mega.nz/folder/4oVRHQaA#vE3Lplfetc9z9SJOxZD6nA 2X_KcjpunkAnime_2.0_Lite_196496_G]
 +
| KCJPUNK
 +
| 2x
 +
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 +
| ESRGAN-lite
 +
| Digital Animation
 +
| 2021-03-23
 +
| Up-scaling Digital Animation
 +
| This is my first attempt to make Light model so I started with 2x version. This model is much faster and give better results than my previous one.
 +
|
 +
|-
  
===Image Restoration===
+
| [https://mega.nz/file/XWpFzRID#398qZx763bMSBsPB2MfgZqROGxdB0VldrIgH_prwa7U 2x_DigitoonLite_216k]
====Image Compression====
+
| SaurusX
=====JPEG Artifacts=====
+
| 2x
{| class="wikitable sortable"
+
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 +
| ESRGAN-lite
 +
| Digital Animation
 +
| 2022-07-05
 +
| Meant as a versatile model for upscaling high detail digital anime and cartoons. Has debanding, MPEG-2 correction, and halo reduction. Trained to handle both 4:3 and 16:9 DVD material with equal efficacy. Will retain a lot of textures except for the really high freq stuff.
 +
| 2x_DigiGradients_Lite_486k.pth
 +
| https://imgsli.com/MTE1Mzg4
 
|-
 
|-
! Model Name
 
! Author
 
! Scale
 
! License
 
! Architecture
 
! Purpose (short)
 
! Date Posted
 
! Purpose (Full)
 
! Pretrained_Model_G
 
! sample
 
  
|-
+
| [https://mega.nz/file/PfhxWCRT#R59B_ovEwdiojZbas76SCbUSv-uNd1ZSOdZ0vQp-6w4 2x_DigiGradients_Lite_486k]
|[https://drive.google.com/drive/folders/1HWokMsUwsR_Mw-NOJgc8OdS8mkdIje_C JPG (00-20%)]
+
| SaurusX
| |[[User:BlueAmulet|BlueAmulet]]
+
| 2x
| 1x
+
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
|[https://creativecommons.org/licenses/by-nc/4.0/ CC BY-NC 4.0]
+
| ESRGAN-lite
| ESRGAN
+
| Digital Animation
| JPEG
+
| 2022-08-24
|  
+
| A very focused model meant for upscaling the TMNT 2003 DVDs. Degradations were added via AVISynth in order to match the video on the TMNT 2003 DVDs to correct the source problems. Problems corrected include aliased red chroma, chroma vertical blur, bad deinterlacing, banding, compression "grain", and poor animation line detail. The AVS scripts for the LR's were run through HCEnc to get authentic low bit rate MPEG-2 artifacts for fixing. By design, the final model gives a very digital-looking result and does not do a good job of retaining textures as the style of TMNT 2003 is all flats and gradients.
|  
+
| none
| Trained on Custom (CC0 Textures)
+
| https://imgsli.com/MTAxNjc1
| JPEG models by Alsa
+
 
 +
https://imgsli.com/ODQ2NzI
 
|-
 
|-
  
|-
+
| [https://mega.nz/file/lVUDRQYa#LVyZdRS1c1Vo7IQcDA9ivBplLqPYPWjl39px7mBKuFo 2x_DigitalFlim_SuperUltraCompact]
|[https://drive.google.com/drive/folders/1HWokMsUwsR_Mw-NOJgc8OdS8mkdIje_C JPG (20-40%)]
+
| Zarxrax
| |[[User:BlueAmulet|BlueAmulet]]
+
| 2x
| 1x
+
|[http://www.wtfpl.net WTFPL]
|[https://creativecommons.org/licenses/by-nc/4.0/ CC BY-NC 4.0]
+
| Compact
| ESRGAN
+
[nf:24 nc:8]
| JPEG
+
| Animation
 +
| 2022-05-14
 +
| This was trained on the dataset that OptimusPrimal used for his DigitalFilm models. This model cleans up the image removing some noise while upscaling. This model is very fast, running around 200x the speed of a standard ESRGAN model on my system. This is primarily a proof of concept for how fast Real-ESRGAN models can get while still producing nice results.
 
|  
 
|  
|  
+
| https://imgsli.com/MTA3ODYx
| Trained on Custom (CC0 Textures)
+
 
| JPEG models by Alsa
+
https://imgsli.com/MTA3ODY0
|-
 
  
 +
https://imgsli.com/MTA3ODY2
  
|[https://drive.google.com/drive/folders/1HWokMsUwsR_Mw-NOJgc8OdS8mkdIje_C JPG (40-60%)]
+
https://imgsli.com/MTA3ODY3
| |[[User:BlueAmulet|BlueAmulet]]
 
| 1x
 
|[https://creativecommons.org/licenses/by-nc/4.0/ CC BY-NC 4.0]
 
| ESRGAN
 
| JPEG
 
|
 
|
 
| Trained on Custom (CC0 Textures)
 
| JPEG models by Alsa
 
 
|-
 
|-
 +
|}
 +
 +
====Manga====
 +
{| class="wikitable sortable"
 +
|-
 +
! Model Name
 +
! Author
 +
! Scale
 +
! License
 +
! Architecture
 +
! Purpose (short)
 +
! Date Posted
 +
! Purpose (Full)
 +
! Pretrained_Model_G
 +
! Sample
  
|[https://drive.google.com/drive/folders/1HWokMsUwsR_Mw-NOJgc8OdS8mkdIje_C JPG (60-80%)]
 
| |[[User:BlueAmulet|BlueAmulet]]
 
| 1x
 
|[https://creativecommons.org/licenses/by-nc/4.0/ CC BY-NC 4.0]
 
| ESRGAN
 
| JPEG
 
|
 
|
 
| Trained on Custom (CC0 Textures)
 
| JPEG models by Alsa
 
 
|-
 
|-
 
+
| [https://drive.google.com/file/d/1uraeNuUPx46HKZ77xUI1PCBBoH-LhikY/view?usp=sharing FSMangaV2]
 
+
| Jacob_
| [https://drive.google.com/drive/folders/1HWokMsUwsR_Mw-NOJgc8OdS8mkdIje_C JPG (80-100%)]
+
| 4x
| |[[User:BlueAmulet|BlueAmulet]]
+
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
| 1x
 
|[https://creativecommons.org/licenses/by-nc/4.0/ CC BY-NC 4.0]
 
 
| ESRGAN
 
| ESRGAN
| JPEG
+
| Manga
|  
+
| 2020-12-14
|  
+
| Manga-style images with or without dithering - cartoons, maybe pixel art, etc
| Trained on Custom (CC0 Textures)
+
| RRDB_PSNR_x4
| JPEG models by Alsa
+
| https://i.imgur.com/oVBVthF.png
 
|-
 
|-
  
 +
|-
 +
| [https://cdn.discordapp.com/attachments/450036622940045325/936099965300768808/4x_eula_digimanga_bw_v1_860k.pth 4x_eula_digimanga_bw_v1_860k]
 +
| end user license agreement#9756
 +
| 4x
 +
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 +
| ESRGAN
 +
| Manga
 +
| 2022-1-26
 +
| Black and white digital manga with halftones.
 +
| 4xPSNR
 +
| [https://slow.pics/c/lp9O3GnZ Clean]
  
|[https://e.pcloud.link/publink/show?code=XZvmsZo4qORoatgHXm4A1CooSUnuy5YD5X JPG (00-20%)]
+
[https://slow.pics/c/AxmHpF3m Heavy Degradation]
| |[[User:Alsa|Alsa]]
 
| 1x
 
|[https://creativecommons.org/licenses/by-nc-sa/4.0/  CC BY-NC-SA 4.0]
 
| ESRGAN
 
| JPEG
 
|
 
|
 
| Trained on Custom (Photos / Manga)
 
|
 
 
|-
 
|-
  
 
+
|-
|[https://e.pcloud.link/publink/show?code=XZimsZmfeBi2HJp9H00wvCRISpnuSRrjgy JPG (20-40%)]
+
| [https://cdn.discordapp.com/attachments/894310505622155264/1008498310597857290/4x_eula_digimanga_bw_v2_nc1_307k.pth 4x_eula_digimanga_bw_v2_nc1_307k]
| |[[User:Alsa|Alsa]]
+
| end user license agreement#9756
| 1x
+
| 4x
|[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]
 
| ESRGAN
 
| ESRGAN
| JPEG
+
| Manga
|  
+
| 2022-08-17
|  
+
|Vast improvement over v1 in low frequency detail; moiré and artifacting reduced significantly and less random noise from JPEG artefacts in the input. Also now only works on 1 channel images, so it runs slightly faster on average and resulting images are much smaller but it might not work on some ESRGAN implementations, I personally recommend using chaiNNer. v1 may still be better in some edge cases.
| Trained on Custom (Photos / Manga)
+
There's also a supplementary 1x model that denoises very low quality LRs and smooths halftones so the image works better with the 4x model. Only trained it to help build the dataset and it's useless for already decent-ish LRs but may help you in some situations.
|  
+
| 4x_eula_digimanga_bw_v1_860k
 +
| https://slow.pics/c/Asbu0xgz
 
|-
 
|-
  
 +
| [https://mega.nz/file/nR51EYYC#RcAKkxb8hAzvZAj4TLxnAukd9gbrYuUeyjUp2l1ffb8 MangaScaleV3]
 +
| Bunzero++
 +
| 2x
 +
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 +
| ESRGAN+
 +
| Manga
 +
| 2021-05-24
 +
| To upscale manga including halftones, instead of trying to smooth them out.
 +
| unreleased model
 +
| https://imgsli.com/NTUyMjg
 +
|-
  
|[https://e.pcloud.link/publink/show?code=XZrmsZbOaINRer4xRn6UWBtGgRK5r7biWy JPG (40-60%)]
+
| [http://www.mediafire.com/file/w3jujtm752hvdj1/Manga109Attempt.pth.zip/file Manga109Attempt]
| |[[User:Alsa|Alsa]]
+
| Kingdomakrillic
| 1x
+
| 4x
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
+
| [https://creativecommons.org/licenses/by/4.0/ CC BY 4.0]
 
| ESRGAN
 
| ESRGAN
| JPEG
+
| Anime / Manga
|  
+
|
|  
+
| Pretrained model: [https://drive.google.com/drive/folders/1ldwajXL50uC7PCS63B4Wato6Dnk-svNL 4xPSNR]
| Trained on Custom (Photos / Manga)
 
 
|  
 
|  
 +
| Manga109
 
|-
 
|-
 +
|}
  
  
|[https://e.pcloud.link/publink/show?code=XZcmsZASeN84hQXCLfhYzUogomG5EdkwPV JPG (60-80%)]
+
====Drawings====
| |[[User:Alsa|Alsa]]
+
 
| 1x
+
{| class="wikitable sortable"
|[https://creativecommons.org/licenses/by-nc-sa/4.0/  CC BY-NC-SA 4.0]
 
| ESRGAN
 
| JPEG
 
|
 
|
 
| Trained on Custom (Photos / Manga)
 
|  
 
 
|-
 
|-
 +
! Model Name
 +
! Author
 +
! Scale
 +
! License
 +
! Architecture
 +
! Purpose (short)
 +
! Date Posted
 +
! Purpose (Full)
 +
! Pretrained_Model_G
 +
! Sample
  
  
| [https://e.pcloud.link/publink/show?code=XZomsZwGOr2Jb4lLXGopIwu8UtNQqpuqyX JPG (80-100%)]
+
|-
| |[[User:Alsa|Alsa]]
+
| [https://www.dropbox.com/s/52hsqa7ymt5pgga/KDM003_scans_1x.pth KDM003_scans_1x.pth]
 +
| BoxDrop
 
| 1x
 
| 1x
|[https://creativecommons.org/licenses/by-nc-sa/4.0/  CC BY-NC-SA 4.0]
+
|  
 
| ESRGAN
 
| ESRGAN
| JPEG
+
| Art
|  
+
| 2019-07-29
|  
+
| Clean up model for scanned illustrations. - Made to remove moire patterns, reduce small imperfections, and correct mild compression artifacts in scanned Kingdom Death: Monster illustrations. CMYK printing often shifts colors, so this is intended to reverse that color shifting as well.
| Trained on Custom (Photos / Manga)
+
| Failed attempts based on ESRGAN_1x_JPEG_80to100
 
|  
 
|  
 
|-
 
|-
  
| [https://mega.nz/file/jcZQ1J6A#qLlSGBFQPWyQ0kUbOcvPN5kEYalPGUu9iR6EkEuS-T4 Kim2091_DeJpeg_v0]
+
|-
| |[[User:Kim2091|Kim2091]]
+
| [https://drive.google.com/open?id=103MX2bvd3GW0MYpC53ABU5VYY6v1t99Y DigiPaint]
| 1x
+
| Rastrum
| [https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
+
| 4x
| ESRGAN Superlite
+
| CC0
| JPEG
+
| ESRGAN
 +
| Art
 +
| 2019-09-07
 +
| Digital Art Upscaler
 +
| 4xfalcoon300(manga).pth
 
|  
 
|  
| Model I forgot to release. This doesn't totally remove JPEG artifacts, but it does a decent job at a fast rate. It seemingly does a better job of retaining detail than some other JPEG models. The model is incomplete, I need to train it further on compact rather than ESRGAN. This is just a temporary release
 
| Custom JPEG dataset
 
| https://cdn.discordapp.com/attachments/903415274521374750/1041539960672632832/1668392853.8196685.png https://cdn.discordapp.com/attachments/903415274521374750/1041540021175463956/1668392885.0326388.png
 
 
|-
 
|-
 +
|}
  
| [https://icedrive.net/1/43GNBihZyi 4x NMKD-PatchySharp]
+
====Cel Animation====
| nmkd
+
{| class="wikitable sortable"
| 4x
 
| WTFPL
 
| ESRGAN (Old Arch)
 
| Art/CGI
 
| 2020-09-12
 
| Upscaler for clean images or images with compression artifacts (jpeg quality >75) - Produces very sharp lines/edges due to NN-Filtered HR images. Proven to produce very, very good results on drawings (sharp lines) and CGI, but should also work pretty well for real-world images.
 
| 4xESRGAN
 
| https://i.imgur.com/PzFkmAg.png
 
 
|-
 
|-
 
+
! Model Name
 
+
! Author
| [https://drive.google.com/open?id=1HWokMsUwsR_Mw-NOJgc8OdS8mkdIje_C 1x_JPEG (by compression level)]
+
! Scale
| |[[User:BlueAmulet|BlueAmulet]]
+
! License
| 1x
+
! Architecture
|[https://creativecommons.org/licenses/by-nc/4.0/ CC BY-NC 4.0]
+
! Purpose (short)
 +
! Date Posted
 +
! Purpose (Full)
 +
! Pretrained_Model_G
 +
! Sample
 +
 
 +
|-
 +
| [https://cdn.discordapp.com/attachments/579685650824036387/812414843939455036/2X_DigitalFilmV5_Lite.pth DigitalFilmV5 Lite]
 +
| OptimusPrimal
 +
| 2x
 +
| WTFPL
 
| ESRGAN
 
| ESRGAN
| JPEG
+
| Traditional Animation
| 2019-08-01
+
| 2021-02-19
| A collection of models for jpeg artifact removal, examples are included in the provided link
+
| Upscaling Dragon Ball Z DBox DVD's.  - grainy sources. It keeps some grain, but also doing some cleaning, sharpening, and fixing.
| JPEG models by |[[User:Alsa|Alsa]]
+
| none
| https://drive.google.com/open?id=1HWokMsUwsR_Mw-NOJgc8OdS8mkdIje_C
+
| https://imgsli.com/NDE0MzQ
|-
+
 
 +
https://imgsli.com/NDE0MzU
 +
 
 +
https://imgsli.com/NDE0MzY
 +
 
 +
https://imgsli.com/NDE0Mzc
  
 +
https://imgsli.com/NDE0Mzg
  
| [https://de-next.owncube.com/index.php/s/w82HLrLWmWi4SQ5 DeJpeg Fatality_PlusULTRA!]
+
https://imgsli.com/NDE0Mzk
| |[[User:Twittman|Twittman]]
 
| 1x
 
| [https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0] (?)
 
| ESRGAN
 
| JPEG
 
| 2019-10-29
 
|
 
| 1x_DeJpeg_Fatality_01_200000_G.pth
 
|
 
 
|-
 
|-
  
 
+
| [https://mega.nz/file/GFFkSL5R#C-dDxaxTZpSPUh_heAUw_idGIF8Xizdr0ktyGi05yFk 4x_HellinaCel]
| [https://mega.nz/folder/vBgGgA5Z#qXsikSiIU8Edemaju91zDA DeCompress]
+
| VGDCKeroro
| |[[User:Kim2091|Kim2091]]
 
 
| 4x
 
| 4x
| [https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
+
| fuckin do whatever
 
| ESRGAN
 
| ESRGAN
| JPEG (works on many compression formats)
+
| Traditional Animation
| 2021-09-07
+
| 2021-03-21
| '''USE 4x-UltraSharp INSTEAD''' - This is UniScaleV2, but intended for images with compression. Seems to work best on realistic images.
+
| A rougher alternative to 4xCelFrames with a focus on realistic looking cels over nice looking cels. It's trained on DetoriationFrames LRs, so if you give it an image straight from the source it will go twice as hard on it. This can be used to your advantage, though. I recommend cleaning it up before running DetoriationFrames, or else it will come out rough.
 
| 4xESRGAN
 
| 4xESRGAN
| https://cdn.discordapp.com/attachments/884864597310459996/884952891721383936/unknown.png https://cdn.discordapp.com/attachments/884864597310459996/884957054807179274/unknown.png https://cdn.discordapp.com/attachments/884864597310459996/884960697883193354/unknown.png https://cdn.discordapp.com/attachments/884864597310459996/884953810097827850/unknown.png
+
| https://imgsli.com/NDU2MjU
 
|-
 
|-
 +
|}
  
  
| [https://de-next.owncube.com/index.php/s/w82HLrLWmWi4SQ5 JPG PlusULTRA (1x_jpg_PlusULTRA_130000.pth)]
+
===Image Restoration===
| |[[User:Twittman|Twittman]]
+
====Image Compression====
 +
=====JPEG Artifacts=====
 +
{| class="wikitable sortable"
 +
|-
 +
! Model Name
 +
! Author
 +
! Scale
 +
! License
 +
! Architecture
 +
! Purpose (short)
 +
! Date Posted
 +
! Purpose (Full)
 +
! Pretrained_Model_G
 +
! Sample
 +
 
 +
|-
 +
|[https://drive.google.com/drive/folders/1HWokMsUwsR_Mw-NOJgc8OdS8mkdIje_C JPG (00-20%)]
 +
| |[[User:BlueAmulet|BlueAmulet]]
 
| 1x
 
| 1x
| [https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0] (?)
+
|[https://creativecommons.org/licenses/by-nc/4.0/ CC BY-NC 4.0]
 
| ESRGAN
 
| ESRGAN
 
| JPEG
 
| JPEG
| 2019-10-29
 
 
|  
 
|  
| 1x_DeJpeg_Fatality_01_200000_G.pth
 
 
|  
 
|  
 +
| Trained on Custom (CC0 Textures)
 +
| JPEG models by Alsa
 
|-
 
|-
  
 
+
|-
| [https://f002.backblazeb2.com/file/ESRGAN/_RELEASE/1x_Saiyajin_DeJPEG_300000_G.pth SaiyaJin DeJpeg]
+
|[https://drive.google.com/drive/folders/1HWokMsUwsR_Mw-NOJgc8OdS8mkdIje_C JPG (20-40%)]
| |[[User:Twittman|Twittman]]
+
| |[[User:BlueAmulet|BlueAmulet]]
 
| 1x
 
| 1x
| [https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0] (?)
+
|[https://creativecommons.org/licenses/by-nc/4.0/ CC BY-NC 4.0]
 
| ESRGAN
 
| ESRGAN
 
| JPEG
 
| JPEG
| 2019-10-29
 
| Dataset: Anime, real life, manga, generated
 
| 1x_DeJpeg_Fatality_PlusULTRA_200000_G.pth
 
 
|  
 
|  
 +
|
 +
| Trained on Custom (CC0 Textures)
 +
| JPEG models by Alsa
 
|-
 
|-
  
| [https://drive.google.com/file/d/1C0Lnn3PfXrm4UX9trjQIWsmxtspjQjmm/view?usp=drivesdk 1x_JPEGDestroyer]
+
 
| |[[User:BlackScout|BlackScout]]
+
|[https://drive.google.com/drive/folders/1HWokMsUwsR_Mw-NOJgc8OdS8mkdIje_C JPG (40-60%)]
 +
| |[[User:BlueAmulet|BlueAmulet]]
 
| 1x
 
| 1x
| [https://www.gnu.org/licenses/gpl-3.0.en.html GNU GPLv3]
+
|[https://creativecommons.org/licenses/by-nc/4.0/ CC BY-NC 4.0]
 
| ESRGAN
 
| ESRGAN
 
| JPEG
 
| JPEG
| 2020-02-19
 
| This model is meant to reduce or eliminate JPEG Compression without making the original images too smooth or killing detail. It manages to do a fairly good job but don't expect overly compressed images to work with this.
 
| 1xESRGAN + previous attempts
 
 
|  
 
|  
 +
|
 +
| Trained on Custom (CC0 Textures)
 +
| JPEG models by Alsa
 
|-
 
|-
  
| [https://drive.google.com/file/d/17k72r01xTkeOzqsaYCN6ZqnAktKqTLyw/view?usp=sharing 4xJaypeg90]
+
|[https://drive.google.com/drive/folders/1HWokMsUwsR_Mw-NOJgc8OdS8mkdIje_C JPG (60-80%)]
| Jacob
+
| |[[User:BlueAmulet|BlueAmulet]]
| 4x
+
| 1x
| [https://www.gnu.org/licenses/gpl-3.0.en.html GNU GPLv3]
+
|[https://creativecommons.org/licenses/by-nc/4.0/ CC BY-NC 4.0]
 
| ESRGAN
 
| ESRGAN
| JPEG compression
+
| JPEG
| 2020-07-20
 
| Photos/realistic 3D with JPEG compression, quality 85-95 and 4:2:0 chroma subsampling. - Created for Myst3 images, since all have 4:2:0 chroma subsampling and existing JPEG models did not give good results. Favors smoothing over over-sharpening.
 
| 4xESRGAN
 
| https://i.imgur.com/n09m2Rt.png
 
|-
 
 
 
| [https://u.pcloud.link/publink/show?code=kZvfeFXZMyUSYScxAizUmTKjax0JQzTk9Gfk 1x_SBDV-DeJPEG-Lite]
 
| Joey
 
| 1x
 
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 
| ESRGAN (lite)
 
| JPEG compression
 
| 2020-10-10
 
 
|  
 
|  
| 1x_DIV2K-Lite_80k.pth
 
 
|  
 
|  
 +
| Trained on Custom (CC0 Textures)
 +
| JPEG models by Alsa
 
|-
 
|-
  
| [https://icedrive.net/1/43GNBihZyi NMKD Jaywreck3 & Jaywreck3-Soft (Lite)]
+
 
| |[[User:nmkd|Nmkd]]
+
| [https://drive.google.com/drive/folders/1HWokMsUwsR_Mw-NOJgc8OdS8mkdIje_C JPG (80-100%)]
 +
| |[[User:BlueAmulet|BlueAmulet]]
 
| 1x
 
| 1x
| WTFPL
+
|[https://creativecommons.org/licenses/by-nc/4.0/ CC BY-NC 4.0]
| ESRGAN Lite [nf=32 nb=12]
+
| ESRGAN
| JPEG compression
+
| JPEG
| 2020-10-13
+
|  
| 1xESRGAN
 
| 1xESRGAN
 
 
|  
 
|  
 +
| Trained on Custom (CC0 Textures)
 +
| JPEG models by Alsa
 
|-
 
|-
  
| [https://mega.nz/folder/fAg21TRI#kBk0synXgNqTGaXiW6nSZA MangaJPEG]
+
 
| Bunzero++
+
|[https://e.pcloud.link/publink/show?code=XZvmsZo4qORoatgHXm4A1CooSUnuy5YD5X JPG (00-20%)]
 +
| |[[User:Alsa|Alsa]]
 
| 1x
 
| 1x
|[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]
 
| ESRGAN
 
| ESRGAN
 
| JPEG
 
| JPEG
| 2021-06-17
+
|  
| Remove JPEG artifacts from manga without destroying screentone and other details
+
|  
| none
+
| Trained on Custom (Photos / Manga)
| https://cdn.discordapp.com/attachments/579685650824036387/855269193762865152/Example_LQ_Input-comparison.png
+
|  
|}
 
 
 
=====Aliasing=====
 
{| class="wikitable sortable"
 
|-
 
! Model Name
 
! Author
 
! Scale
 
! License
 
! Architecture
 
! Purpose (short)
 
! Date Posted
 
! Purpose (Full)
 
! Pretrained_Model_G
 
! sample
 
 
|-
 
|-
  
|-
+
 
| [https://drive.google.com/open?id=11nMqVF9sUN2jRRagYc3wdWNX5acEy0MJ SS Anti Alias 9x]
+
|[https://e.pcloud.link/publink/show?code=XZimsZmfeBi2HJp9H00wvCRISpnuSRrjgy JPG (20-40%)]
| |[[User:BlueAmulet|BlueAmulet]]
+
| |[[User:Alsa|Alsa]]
 
| 1x
 
| 1x
| [https://creativecommons.org/licenses/by-nc/4.0/ CC BY-NC 4.0]
+
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 
| ESRGAN
 
| ESRGAN
| Anti aliasing
+
| JPEG
 +
|
 
|  
 
|  
| Dataset: Custom (9x Supersampling AA)
+
| Trained on Custom (Photos / Manga)  
| none (no interpolation)  
 
 
|  
 
|  
 
|-
 
|-
  
|-
+
 
| [https://de-next.owncube.com/index.php/s/mmYCcK4H3rQQR5Y Anti Aliasing]
+
|[https://e.pcloud.link/publink/show?code=XZrmsZbOaINRer4xRn6UWBtGgRK5r7biWy JPG (40-60%)]
| |[[User:Twittman|Twittman]]
+
| |[[User:Alsa|Alsa]]
 
| 1x
 
| 1x
| [https://creativecommons.org/licenses/by-nc/4.0/ CC BY-NC 4.0]
+
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 
| ESRGAN
 
| ESRGAN
| Anti aliasing / Images with pixelated edges
+
| JPEG
 +
|
 
|  
 
|  
| Dataset: Custom (?)
+
| Trained on Custom (Photos / Manga)  
| none (no interpolation)  
 
 
|  
 
|  
 
|-
 
|-
  
|-
+
 
| [https://mega.nz/folder/yg0lHQoJ#sP8_BfDk2YlshFjOL9Qrtg 1x_GainRESV3 (Aggro,Natural,Passive)]
+
|[https://e.pcloud.link/publink/show?code=XZcmsZASeN84hQXCLfhYzUogomG5EdkwPV JPG (60-80%)]
| CF2lter
+
| |[[User:Alsa|Alsa]]
 
| 1x
 
| 1x
| WTFPL
+
|[https://creativecommons.org/licenses/by-nc-sa/4.0/  CC BY-NC-SA 4.0]
 
| ESRGAN
 
| ESRGAN
| Anti aliasing / Deblur
+
| JPEG
| 2022-02-26
+
|
| To eliminate aliasing and general artifacts caused by not enough resolution while bringing out details Im stopping its training here because it's getting worse, i think of some aligment issues by game's rendering pipeline + downscaling... Dataset: 5K resolution shots from paladins rendered by 200% for hr and 37.5%(1080p) for lr then downscaled to 1080
+
|  
| BCGONE_DetailedV2
+
| Trained on Custom (Photos / Manga)  
| https://cdn.discordapp.com/attachments/579685650824036387/947132730313949184/cmp2.png https://cdn.discordapp.com/attachments/579685650824036387/947132731056324668/cmp.png
+
|  
 
|-
 
|-
  
|}
 
  
=====GIF=====
+
| [https://e.pcloud.link/publink/show?code=XZomsZwGOr2Jb4lLXGopIwu8UtNQqpuqyX JPG (80-100%)]
{| class="wikitable sortable"
+
| |[[User:Alsa|Alsa]]
|-
+
| 1x
! Model Name
+
|[https://creativecommons.org/licenses/by-nc-sa/4.0/  CC BY-NC-SA 4.0]
! Author
 
! Scale
 
! License
 
! Architecture
 
! Purpose (short)
 
! Date Posted
 
! Purpose (Full)
 
! Pretrained_Model_G
 
! sample
 
|-
 
| [https://icedrive.net/1/43GNBihZyi DeGif]
 
| |[[User:nmkd|nmkd]]
 
| 2x
 
| WTFPL
 
 
| ESRGAN
 
| ESRGAN
| GIF Restoration
+
| JPEG
 +
|
 
|  
 
|  
| GIF Restoration
+
| Trained on Custom (Photos / Manga)
| [https://mega.nz/#!vtgSWKQT!K7Asn2zKe4N70R2aV89KEMTKhH3aiyGAAiuQDJF09qs 2xESRGAN]
 
 
|  
 
|  
 
|-
 
|-
|}
 
  
=====DDS (BC1/DXT1, BC3/DXT5 Compression)=====
+
| [https://mega.nz/file/jcZQ1J6A#qLlSGBFQPWyQ0kUbOcvPN5kEYalPGUu9iR6EkEuS-T4 Kim2091_DeJpeg_v0]
{| class="wikitable sortable"
+
| |[[User:Kim2091|Kim2091]]
 +
| 1x
 +
| [https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 +
| ESRGAN Superlite
 +
| JPEG
 +
|
 +
| Model I forgot to release. This doesn't totally remove JPEG artifacts, but it does a decent job at a fast rate. It seemingly does a better job of retaining detail than some other JPEG models. The model is incomplete, I need to train it further on compact rather than ESRGAN. This is just a temporary release
 +
| Custom JPEG dataset
 +
| [https://cdn.discordapp.com/attachments/903415274521374750/1041539960672632832/1668392853.8196685.png Sample 1]
 +
 
 +
[https://cdn.discordapp.com/attachments/903415274521374750/1041540021175463956/1668392885.0326388.png Sample 2]
 
|-
 
|-
! Model Name
 
! Author
 
! Scale
 
! License
 
! Architecture
 
! Purpose (short)
 
! Date Posted
 
! Purpose (Full)
 
! Pretrained_Model_G
 
! sample
 
  
|-
+
| [https://icedrive.net/1/43GNBihZyi 4x NMKD-PatchySharp]
| [https://drive.google.com/open?id=1LHplsPRqhmjR28jGgRlEekeP_bvx3nUC 1x_BC1-smooth2.pth]
+
| nmkd
 +
| 4x
 +
| WTFPL
 +
| ESRGAN (Old Arch)
 +
| Art/CGI
 +
| 2020-09-12
 +
| Upscaler for clean images or images with compression artifacts (jpeg quality >75) - Produces very sharp lines/edges due to NN-Filtered HR images. Proven to produce very, very good results on drawings (sharp lines) and CGI, but should also work pretty well for real-world images.
 +
| 4xESRGAN
 +
| https://i.imgur.com/PzFkmAg.png
 +
|-
 +
 
 +
 
 +
| [https://drive.google.com/open?id=1HWokMsUwsR_Mw-NOJgc8OdS8mkdIje_C 1x_JPEG (by compression level)]
 
| |[[User:BlueAmulet|BlueAmulet]]
 
| |[[User:BlueAmulet|BlueAmulet]]
 
| 1x
 
| 1x
 
|[https://creativecommons.org/licenses/by-nc/4.0/ CC BY-NC 4.0]
 
|[https://creativecommons.org/licenses/by-nc/4.0/ CC BY-NC 4.0]
 
| ESRGAN
 
| ESRGAN
| BC1 Compression
+
| JPEG
| 2019-09-05
+
| 2019-08-01
| A model to help remove compression artifacts in BC1-BC3/DXT1-DXT5 compressed images (these all have color encoded the same way)
+
| A collection of models for jpeg artifact removal, examples are included in the provided link
| none
+
| JPEG models by |[[User:Alsa|Alsa]]
|  
+
| [https://drive.google.com/open?id=1HWokMsUwsR_Mw-NOJgc8OdS8mkdIje_C Sample Gallery]
 
|-
 
|-
  
|-
+
 
| [https://drive.google.com/drive/folders/1y-QoWGAnF8YiX764rWrDOGzndWtpCfF3?usp=sharing x1_ITF_SkinDiffDDS_v1]
+
| [https://de-next.owncube.com/index.php/s/w82HLrLWmWi4SQ5 DeJpeg Fatality_PlusULTRA!]
| intheflesh#3116
+
| |[[User:Twittman|Twittman]]
 
| 1x
 
| 1x
|[https://creativecommons.org/licenses/by-nc/4.0/ CC BY-NC 4.0]
+
| [https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0] (?)
 
| ESRGAN
 
| ESRGAN
| BC1 Compression
+
| JPEG
| 2022-08-24
+
| 2019-10-29
| Removes banding, blocking, dithering, aliasing, noise and color tint on DDS Compressed Skin Diffuse Textures.  
+
|  
 +
| 1x_DeJpeg_Fatality_01_200000_G.pth
 +
|
 +
|-
 +
 
  
This should work extremely well on most modern DDS compression types. The training set was compressed with BC3/DXT5, BC3/DXT5 Fast, BC2/DXT3, BC2/DXT3 Fast, and a small number of JPEG compressed images to cover outliers.
+
| [https://mega.nz/folder/vBgGgA5Z#qXsikSiIU8Edemaju91zDA DeCompress]
 +
| |[[User:Kim2091|Kim2091]]
 +
| 4x
 +
| [https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 +
| ESRGAN
 +
| JPEG (works on many compression formats)
 +
| 2021-09-07
 +
| '''USE 4x-UltraSharp INSTEAD''' - This is UniScaleV2, but intended for images with compression. Seems to work best on realistic images.
 +
| 4xESRGAN
 +
| [https://cdn.discordapp.com/attachments/884864597310459996/884952891721383936/unknown.png Sample 1]
  
This model is trained to remove the slight green color tint that DDS compression tends to add to skin textures, so the model output will not match the original color tone of the input image. This is the desired result though, as DDS compression shifts the colors to a sickly green tint and this model corrects that to more natural color tones.
+
[https://cdn.discordapp.com/attachments/884864597310459996/884957054807179274/unknown.png Sample 2]
  
The training set included faces, body parts, eyes, mouths and hair in a variety of skin types and tones so it should work well on most related diffuse textures.
+
[https://cdn.discordapp.com/attachments/884864597310459996/884960697883193354/unknown.png Sample 3]
  
However it's not just limited to skin, many other images and textures can be cleaned with this model. Designed to be used as a first step cleaning pass before applying additional models after. Check out the other ITF Models.
+
[https://cdn.discordapp.com/attachments/884864597310459996/884953810097827850/unknown.png Sample 4]
| 1x_BC1-smooth2.pth
 
| https://imgsli.com/MTIyMzE5 https://imgsli.com/MTIyMzIw https://imgsli.com/MTIyMzIx
 
 
|-
 
|-
  
|-
+
 
| [https://drive.google.com/file/d/1WXBNdlqDWV10a3L1_U7zuNkSpN9aBF0M/view?usp=sharing 1x_DXTDecompressor_Source_V3-300000_G]
+
| [https://de-next.owncube.com/index.php/s/w82HLrLWmWi4SQ5 JPG PlusULTRA (1x_jpg_PlusULTRA_130000.pth)]
| JosephtheKP
+
| |[[User:Twittman|Twittman]]
 
| 1x
 
| 1x
| WTFPL
+
| [https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0] (?)
 
| ESRGAN
 
| ESRGAN
| DXT1 Compression
+
| JPEG
| 2021-12-21
+
| 2019-10-29
| Removing compression artifacts from DXT1 compressed textures. This model was created to remove DXT1 compression artifacts from textures imported into the Source Engine. Compressed textures in the engine sometimes have a green-tint which this model also corrects. The data for this model contained a good mix of diffuse textures and normal maps which means this model is pretty good at removing compression from normals as well. Creating this model was a real learning experience for me and I hope someone finds a good use for it.
+
|  
|
+
| 1x_DeJpeg_Fatality_01_200000_G.pth
| https://imgsli.com/ODcwNDg https://imgsli.com/ODcwNDk https://imgsli.com/ODcwNTA https://imgsli.com/ODcwNTE
+
|  
 
|-
 
|-
  
|-
+
 
| [https://mega.nz/file/TltS1TQY#CW3zaI24pxFXggeRjgnXb4bSfjnQsY_ch1bogCjMCmU 1x_DEDXT]
+
| [https://f002.backblazeb2.com/file/ESRGAN/_RELEASE/1x_Saiyajin_DeJPEG_300000_G.pth SaiyaJin DeJpeg]
| CF2lter
+
| |[[User:Twittman|Twittman]]
 
| 1x
 
| 1x
| WTFPL
+
| [https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0] (?)
 
| ESRGAN
 
| ESRGAN
| DXT compression
+
| JPEG
| 2022-02-26
+
| 2019-10-29
| To retain details while removing artifacts caused by dxt compression on textures
+
| Dataset: Anime, real life, manga, generated
|
+
| 1x_DeJpeg_Fatality_PlusULTRA_200000_G.pth
| https://cdn.discordapp.com/attachments/579685650824036387/947116367205793882/cmp.png
+
| [https://f002.backblazeb2.com/file/ESRGAN/image_samples/SaiyaJin%20DeJpeg/sample_001.png Example 1]
 +
 
 +
[https://f002.backblazeb2.com/file/ESRGAN/image_samples/SaiyaJin%20DeJpeg/sample_002.png Example 2]
 +
 
 +
[https://f002.backblazeb2.com/file/ESRGAN/image_samples/SaiyaJin%20DeJpeg/sample_006.png Example 3]
 
|-
 
|-
  
|-
+
| [https://drive.google.com/file/d/1C0Lnn3PfXrm4UX9trjQIWsmxtspjQjmm/view?usp=drivesdk 1x_JPEGDestroyer]
| [https://e.pcloud.link/publink/show?code=XZXbsZK03Y0UsmtR0W7wRMHWGvYkzgPII7 1x_artifacts_bc1_free_alsa.pth]
+
| |[[User:BlackScout|BlackScout]]
| |[[User:Alsa|Alsa]]
 
 
| 1x
 
| 1x
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
+
| [https://www.gnu.org/licenses/gpl-3.0.en.html GNU GPLv3]
 
| ESRGAN
 
| ESRGAN
| BC1 Compression
+
| JPEG
|  
+
| 2020-02-19
|  
+
| This model is meant to reduce or eliminate JPEG Compression without making the original images too smooth or killing detail. It manages to do a fairly good job but don't expect overly compressed images to work with this.
| [https://mega.nz/#!r2hEAAAb!zdk-Ka6VCqVnKThtfJVYM0NnVBSUyeqsjYs-NgKjLkc BC1 take 2]
+
| 1xESRGAN + previous attempts
 
|  
 
|  
 
|-
 
|-
  
| [https://drive.google.com/drive/u/1/folders/1pnyPklpLPtMAHeNC5NVU1mV_Qh61wWnx 1x_BCGone_Smooth_110000_G]
+
| [https://drive.google.com/file/d/17k72r01xTkeOzqsaYCN6ZqnAktKqTLyw/view?usp=sharing 4xJaypeg90]
| |[[User:Mutin_Choler|Mutin_Choler]]
+
| Jacob
| 1x
+
| 4x
| WTFPL
+
| [https://www.gnu.org/licenses/gpl-3.0.en.html GNU GPLv3]
 
| ESRGAN
 
| ESRGAN
| BC1 Compression
+
| JPEG compression
| 2020-08-02
+
| 2020-07-20
| Attempts to remove the damages done by BC1 compression.
+
| Photos/realistic 3D with JPEG compression, quality 85-95 and 4:2:0 chroma subsampling. - Created for Myst3 images, since all have 4:2:0 chroma subsampling and existing JPEG models did not give good results. Favors smoothing over over-sharpening.
| 1xESRGAN
+
| 4xESRGAN
| https://imgsli.com/MjU5MzQ/4/3
+
| https://i.imgur.com/n09m2Rt.png
 
|-
 
|-
  
| [https://drive.google.com/drive/folders/1pnyPklpLPtMAHeNC5NVU1mV_Qh61wWnx?usp=sharing 1x_BCGone-DetailedV2_40-60_115000_G]
+
| [https://u.pcloud.link/publink/show?code=kZvfeFXZMyUSYScxAizUmTKjax0JQzTk9Gfk 1x_SBDV-DeJPEG-Lite]
| |[[User:Mutin_Choler|Mutin_Choler]]
+
| Joey
 +
| 1x
 +
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 +
| ESRGAN (lite)
 +
| JPEG compression
 +
| 2020-10-10
 +
|
 +
| 1x_DIV2K-Lite_80k.pth
 +
|
 +
|-
 +
 
 +
| [https://icedrive.net/1/43GNBihZyi NMKD Jaywreck3 & Jaywreck3-Soft (Lite)]
 +
| |[[User:nmkd|Nmkd]]
 
| 1x
 
| 1x
 
| WTFPL
 
| WTFPL
| ESRGAN
+
| ESRGAN Lite [nf=32 nb=12]
| BC1 Compression
+
| JPEG compression
| 2021-03-18
+
| 2020-10-13
|  
+
| 1xESRGAN
 
| 1xESRGAN
 
| 1xESRGAN
 
|  
 
|  
 
|-
 
|-
 +
 +
| [https://mega.nz/folder/fAg21TRI#kBk0synXgNqTGaXiW6nSZA MangaJPEG]
 +
| Bunzero++
 +
| 1x
 +
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 +
| ESRGAN
 +
| JPEG
 +
| 2021-06-17
 +
| Remove JPEG artifacts from manga without destroying screentone and other details
 +
| none
 +
| [https://cdn.discordapp.com/attachments/579685650824036387/855269193762865152/Example_LQ_Input-comparison.png Sample]
 
|}
 
|}
  
=====Dithering=====
+
=====Aliasing=====
 
{| class="wikitable sortable"
 
{| class="wikitable sortable"
 
|-
 
|-
Line 2,234: Line 2,360:
 
! Purpose (Full)
 
! Purpose (Full)
 
! Pretrained_Model_G
 
! Pretrained_Model_G
! sample
+
! Sample
 
+
|-
  
 
|-
 
|-
| [https://e.pcloud.link/publink/show?code=XZ5bsZtd0ldzEF2UFLSYqIBNhqNH21slxX 1x_artifacts_dithering_alsa.pth (pcloud)][https://mega.nz/file/IsABVIJb#ljH6JDvFahi6KLWpaeQ_G9DOSQlvhwjbitlPBbSHcuo (mega)]
+
| [https://drive.google.com/open?id=11nMqVF9sUN2jRRagYc3wdWNX5acEy0MJ SS Anti Alias 9x]
| [[User:Alsa|Alsa]]
+
| |[[User:BlueAmulet|BlueAmulet]]
 
| 1x
 
| 1x
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
+
| [https://creativecommons.org/licenses/by-nc/4.0/ CC BY-NC 4.0]
 
| ESRGAN
 
| ESRGAN
| Dithering
+
| Anti aliasing
 
|  
 
|  
| Dithered Images
+
| Dataset: Custom (9x Supersampling AA)
| [https://mega.nz/#!7io2gSQR!UB9u2k51daixTgC2H0LdOWzNlkDyIDHwxBX4BVY2J3k JPG (0-20%)]
+
| none (no interpolation)  
 
|  
 
|  
 
|-
 
|-
  
 
|-
 
|-
| [https://drive.google.com/file/d/1H4KQyhcknOoExjvDdsoxAgTBMO7JuJ3w/view 4xFSDedither]
+
| [https://de-next.owncube.com/index.php/s/mmYCcK4H3rQQR5Y Anti Aliasing]
| Jacob
+
| |[[User:Twittman|Twittman]]
| 4x
+
| 1x
| GNU GPL3
+
| [https://creativecommons.org/licenses/by-nc/4.0/ CC BY-NC 4.0]
 
| ESRGAN
 
| ESRGAN
| Dithering
+
| Anti aliasing / Images with pixelated edges
| 2020-01-29
+
|  
| 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 (https://github.com/ManuelFritsche/real-world-sr/tree/master/esrgan-fs/codes) for better details compared to plain ESRGAN.
+
| Dataset: Custom (?)  
| RRDB_ESRGAN_x4
+
| none (no interpolation)
| https://imgsli.com/MTI1NTY/
+
|  
 
|-
 
|-
  
 
|-
 
|-
| [https://drive.google.com/open?id=1Vtwt9kgnmju1edibl07WI6Rg7CrAlHIx 4xFSDedither_Manga]
+
| [https://mega.nz/folder/yg0lHQoJ#sP8_BfDk2YlshFjOL9Qrtg 1x_GainRESV3 (Aggro,Natural,Passive)]
| Jacob
+
| CF2lter
| 4x
+
| 1x
| [https://www.gnu.org/licenses/gpl-3.0.en.html GNU GPLv3]
+
| WTFPL
| ESRGAN-FS
+
| ESRGAN
| Dithering
+
| Anti aliasing / Deblur
| 2020-04-21
+
| 2022-02-26
| Cartoons/pixel art/other non-realistic stuff with dithering
+
| To eliminate aliasing and general artifacts caused by not enough resolution while bringing out details Im stopping its training here because it's getting worse, i think of some aligment issues by game's rendering pipeline + downscaling... Dataset: 5K resolution shots from paladins rendered by 200% for hr and 37.5%(1080p) for lr then downscaled to 1080
| RRDB_ESRGAN_x4.pth
+
| BCGONE_DetailedV2
| https://imgsli.com/MTQ3Nzc
+
| [https://cdn.discordapp.com/attachments/579685650824036387/947132730313949184/cmp2.png Sample 1]
 +
 
 +
[https://cdn.discordapp.com/attachments/579685650824036387/947132731056324668/cmp.png Sample 2]
 
|-
 
|-
  
| [https://buildism.net/files/4xFSDedither_Riven.pth 4xFSDedither_Riven]
 
| Jacob
 
| 4x
 
| [https://www.gnu.org/licenses/gpl-3.0.en.html GNU GPLv3]
 
| ESRGAN-FS
 
| Dithering
 
| 2020-07-11
 
| Fine-tuned 4xFSDedither 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.
 
| 4xESRGAN
 
| https://imgsli.com/MTg5NTM https://imgsli.com/MTg5NTQ https://imgsli.com/MTg5NTU https://i.imgur.com/j7Wtn0G.png
 
|-
 
 
| [https://drive.google.com/drive/folders/1XKREcpHnaYz_dcmkMPDbyETcXzDcQsR2?usp=sharing 1x_DitherDeleter-Smooth_104000_G]
 
| |[[User:Mutin_Choler|Mutin_Choler]]
 
| 1x
 
| WTFPL
 
| ESRGAN
 
| Dithering
 
| 2021-01-23
 
| 1xESRGAN
 
| https://imgsli.com/Mzg0MTk/1/2
 
|-
 
 
| [https://drive.google.com/drive/folders/1XKREcpHnaYz_dcmkMPDbyETcXzDcQsR2?usp=sharing 1x_DitherDeleterV3-Smooth]
 
| |[[User:Mutin_Choler|Mutin_Choler]]
 
| 1x
 
| WTFPL
 
| ESRGAN
 
| Dithering
 
| 2021-04-18
 
| Attempts to remove the damages done by dithering. For this model, I downscaled all of the HR images to make their pixel size closer to 1000x1000 using the Box filter and then downscaled them again by 50% using the Point filter. Afterwards, I applied 32-bit Riemersma to every image in the dataset.
 
| 1xESRGAN
 
| https://imgsli.com/NTA5MjU/1/0
 
|-
 
 
|}
 
|}
  
====Blurring====
+
=====GIF=====
 
 
 
{| class="wikitable sortable"
 
{| class="wikitable sortable"
 
|-
 
|-
Line 2,325: Line 2,418:
 
! Purpose (Full)
 
! Purpose (Full)
 
! Pretrained_Model_G
 
! Pretrained_Model_G
! sample
+
! Sample
 
 
 
|-
 
|-
| [https://mega.nz/file/CYoXCAKR#eRKwpP9rCL9vFCsO41oVImvmVwAEOtYUlp2CquX0Fw8 1x_ThePi7on-Solidd_Deborutify_UltraLite_260k_G]
+
| [https://icedrive.net/1/43GNBihZyi DeGif]
| ThePi7on
+
| |[[User:nmkd|nmkd]]
| 1x
+
| 2x
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
+
| WTFPL
 
| ESRGAN
 
| ESRGAN
| Deblur
+
| GIF Restoration
| 2021-05-09
+
|  
| Sharpening, line darkening and slight line thinning, specifically made for the Boruto anime.
+
| GIF Restoration
|
+
| [https://mega.nz/#!vtgSWKQT!K7Asn2zKe4N70R2aV89KEMTKhH3aiyGAAiuQDJF09qs 2xESRGAN]
|
+
|  
 
|-
 
|-
 +
|}
  
 +
=====DDS (BC1/DXT1, BC3/DXT5 Compression)=====
 +
{| class="wikitable sortable"
 +
|-
 +
! Model Name
 +
! Author
 +
! Scale
 +
! License
 +
! Architecture
 +
! Purpose (short)
 +
! Date Posted
 +
! Purpose (Full)
 +
! Pretrained_Model_G
 +
! Sample
  
 
|-
 
|-
| [https://mega.nz/file/zcBRHRbJ#iLStv-F5ax-vQBPG7N6OaRv_xBhRHux3pdnSG43GuMg 1x_DeBLR]
+
| [https://drive.google.com/open?id=1LHplsPRqhmjR28jGgRlEekeP_bvx3nUC 1x_BC1-smooth2.pth]
| BlackScout
+
| |[[User:BlueAmulet|BlueAmulet]]
 
| 1x
 
| 1x
| Unknown or None
+
|[https://creativecommons.org/licenses/by-nc/4.0/ CC BY-NC 4.0]
 
| ESRGAN
 
| ESRGAN
| Deblur
+
| BC1 Compression
| 2020-02-19
+
| 2019-09-05
| General Deblurring
+
| A model to help remove compression artifacts in BC1-BC3/DXT1-DXT5 compressed images (these all have color encoded the same way)
|
+
| none
|
+
|  
 
|-
 
|-
  
| [https://cloud.owncube.com/s/p3w7M2sfaWWYCK9/download?path=%2F&files=1x_mdeblur.pth mdeblur]
+
|-
| LyonHrt
+
| [https://drive.google.com/drive/folders/1y-QoWGAnF8YiX764rWrDOGzndWtpCfF3?usp=sharing x1_ITF_SkinDiffDDS_v1]
 +
| intheflesh#3116
 
| 1x
 
| 1x
|UNKNOWN
+
|[https://creativecommons.org/licenses/by-nc/4.0/ CC BY-NC 4.0]
 
| ESRGAN
 
| ESRGAN
| Blurring
+
| BC1 Compression
|  
+
| 2022-08-24
| Strong deblurring model
+
| Removes banding, blocking, dithering, aliasing, noise and color tint on DDS Compressed Skin Diffuse Textures.
|  
+
 
|  
+
This should work extremely well on most modern DDS compression types. The training set was compressed with BC3/DXT5, BC3/DXT5 Fast, BC2/DXT3, BC2/DXT3 Fast, and a small number of JPEG compressed images to cover outliers.
 +
 
 +
This model is trained to remove the slight green color tint that DDS compression tends to add to skin textures, so the model output will not match the original color tone of the input image. This is the desired result though, as DDS compression shifts the colors to a sickly green tint and this model corrects that to more natural color tones.
 +
 
 +
The training set included faces, body parts, eyes, mouths and hair in a variety of skin types and tones so it should work well on most related diffuse textures.
 +
 
 +
However it's not just limited to skin, many other images and textures can be cleaned with this model. Designed to be used as a first step cleaning pass before applying additional models after. Check out the other ITF Models.
 +
| 1x_BC1-smooth2.pth
 +
| https://imgsli.com/MTIyMzE5
 +
 
 +
https://imgsli.com/MTIyMzIw
 +
 
 +
https://imgsli.com/MTIyMzIx
 
|-
 
|-
  
| [https://1drv.ms/u/s!Aip-EMByJHY27TDhiFPB39jM0pwq?e=5zAFtO 1x_PixelSharpen_100000]
+
|-
| DinJerr
+
| [https://drive.google.com/file/d/1WXBNdlqDWV10a3L1_U7zuNkSpN9aBF0M/view?usp=sharing 1x_DXTDecompressor_Source_V3-300000_G]
 +
| JosephtheKP
 
| 1x
 
| 1x
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
+
| WTFPL
 
| ESRGAN
 
| ESRGAN
| Pixel Art
+
| DXT1 Compression
| 2021-03-19
+
| 2021-12-21
| Restores blurry/upscaled pixel art.
+
| Removing compression artifacts from DXT1 compressed textures. This model was created to remove DXT1 compression artifacts from textures imported into the Source Engine. Compressed textures in the engine sometimes have a green-tint which this model also corrects. The data for this model contained a good mix of diffuse textures and normal maps which means this model is pretty good at removing compression from normals as well. Creating this model was a real learning experience for me and I hope someone finds a good use for it.
| 1xESRGAN
+
|
| https://imgsli.com/NDUxNDQ/5/4
+
| https://imgsli.com/ODcwNDg
 +
 
 +
https://imgsli.com/ODcwNDk
 +
 
 +
https://imgsli.com/ODcwNTA
 +
 
 +
https://imgsli.com/ODcwNTE
 
|-
 
|-
  
| [https://1drv.ms/u/s!Aip-EMByJHY28w0Zkd9fdANYH3j6?e=9zwxII 1x_ArtClarity]
+
|-
| DinnJerr
+
| [https://mega.nz/file/TltS1TQY#CW3zaI24pxFXggeRjgnXb4bSfjnQsY_ch1bogCjMCmU 1x_DEDXT]
 +
| CF2lter
 
| 1x
 
| 1x
 
| WTFPL
 
| WTFPL
 
| ESRGAN
 
| ESRGAN
| Pixel Art
+
| DXT compression
| 2021-08-05
+
| 2022-02-26
| Texture retaining denoiser and sharpener for digital artwork
+
| To retain details while removing artifacts caused by dxt compression on textures
| 4xPSNR
 
 
|
 
|
 +
| [https://cdn.discordapp.com/attachments/579685650824036387/947116367205793882/cmp.png Sample]
 
|-
 
|-
  
| [https://de-next.owncube.com/index.php/s/aAojXwLTPZto8rP 1x_Fatality_DeBlur]
 
| |[[User:Twittman|Twittman]]
 
| 1x
 
| [https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 
| ESRGAN
 
| Deblur
 
| 2019-10-16
 
| Dataset: Mix of anime, manga, and photos
 
| 1x_DeJpeg_Fatality_01_175000_G.pth
 
|
 
 
|-
 
|-
 
+
| [https://e.pcloud.link/publink/show?code=XZXbsZK03Y0UsmtR0W7wRMHWGvYkzgPII7 1x_artifacts_bc1_free_alsa.pth]
| [https://de-next.owncube.com/index.php/s/nmNfcd33mnLLwEC UnResize V3 (1x_UnResize_V3_200000_G.pth)]
+
| |[[User:Alsa|Alsa]]
| |[[User:Twittman|Twittman]]
 
 
| 1x
 
| 1x
| [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]
 
| ESRGAN
 
| ESRGAN
| Deblur / Unresize
+
| BC1 Compression
 +
|
 
|  
 
|  
| Dataset: Anime, manga, and photos -
+
| [https://mega.nz/#!r2hEAAAb!zdk-Ka6VCqVnKThtfJVYM0NnVBSUyeqsjYs-NgKjLkc BC1 take 2]
Purpose: Fix images that have been arbitrarily / poorly resized, such as non-integer nearest-neighbor upscaling/downscaling -
 
Also acts as an image sharpener/deblur when used on slightly soft inputs
 
| 1x_UnResize_MKII_030000_G.pth
 
 
|  
 
|  
 
|-
 
|-
  
| [https://mega.nz/folder/OVomUZyT#vDyNntGY1MirBEACNjpqVg 1x-Focus and 1x-Focus_Moderate]
+
| [https://drive.google.com/drive/u/1/folders/1pnyPklpLPtMAHeNC5NVU1mV_Qh61wWnx 1x_BCGone_Smooth_110000_G]
| |[[User:Kim2091|Kim2091]]
+
| |[[User:Mutin_Choler|Mutin_Choler]]
 
| 1x
 
| 1x
| [https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
+
| WTFPL
 
| ESRGAN
 
| ESRGAN
| Deblur
+
| BC1 Compression
|  
+
| 2020-08-02
| These models deblur most images. It was trained mostly on aniso2 and iso blurring (BSRGAN augmentation) with some gaussian mixed in. It performs well on most blurry images, but I'd recommend using something like Fatality_Deblur for very strong gaussian blur. Dataset: The UniScale Dataset + My Fabric dataset
+
| Attempts to remove the damages done by BC1 compression.
 
| 1xESRGAN
 
| 1xESRGAN
| [https://cdn.discordapp.com/attachments/893483065664483338/893483645803847690/unknown.png Sample 1] [https://cdn.discordapp.com/attachments/893483065664483338/893497416207188028/unknown.png Sample 2] [https://cdn.discordapp.com/attachments/893483065664483338/893483280895193118/unknown.png Sample 3] [https://cdn.discordapp.com/attachments/893483065664483338/893496207832408135/unknown.png Sample 4]
+
| https://imgsli.com/MjU5MzQ/4/3
 
|-
 
|-
  
| [https://de-next.owncube.com/index.php/s/HyygDgXEB45JqBH ReFocus V3 (1x_ReFocus_V3_140000_G.pth)]
+
| [https://drive.google.com/drive/folders/1pnyPklpLPtMAHeNC5NVU1mV_Qh61wWnx?usp=sharing 1x_BCGone-DetailedV2_40-60_115000_G]
| |[[User:Twittman|Twittman]]
+
| |[[User:Mutin_Choler|Mutin_Choler]]
 
| 1x
 
| 1x
| [https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
+
| WTFPL
 
| ESRGAN
 
| ESRGAN
| Deblur / ReFocus
+
| BC1 Compression
| 2022-03-12
+
| 2021-03-18
| Dataset: Photos, Anime, and manga    - 
 
Purpose: DeBlur, ReFocus, Sharpen Real life style images, but will work on Anime images too.
 
| 1x_Saiyajin_DeJPEG_300000_G.pth
 
 
|  
 
|  
|-
+
| 1xESRGAN
 
 
| [https://de-next.owncube.com/index.php/s/GjbBw7pcgm5gmYm ReFocus Cleanly (1x_ReFocus_Cleanly_100000_G.pth)]
 
| |[[User:Twittman|Twittman]]
 
| 1x
 
| [https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 
| ESRGAN
 
| Deblur / ReFocus
 
| 2022-01-21
 
| Dataset: Anime, manga, and photos    - 
 
Purpose: DeBlur, ReFocus, Sharpen Manga, Anime and cartoon style images, but will work on real life images too.
 
| 1x_ReFocus_V3_110000_G.pth
 
 
|  
 
|  
 
|-
 
|-
 
 
|}
 
|}
  
====Banding====
+
=====Dithering=====
 
{| class="wikitable sortable"
 
{| class="wikitable sortable"
 
|-
 
|-
Line 2,468: Line 2,567:
 
! Purpose (Full)
 
! Purpose (Full)
 
! Pretrained_Model_G
 
! Pretrained_Model_G
! sample
+
! Sample
 +
 
  
 
|-
 
|-
| [https://drive.google.com/file/d/1GNguVMV4E5t_4dn23w6UMXVKN6Ff9XPt/view?usp=sharing 1x_N64clean]
+
| [https://e.pcloud.link/publink/show?code=XZ5bsZtd0ldzEF2UFLSYqIBNhqNH21slxX 1x_artifacts_dithering_alsa.pth (pcloud)][https://mega.nz/file/IsABVIJb#ljH6JDvFahi6KLWpaeQ_G9DOSQlvhwjbitlPBbSHcuo (mega)]
| |[[User:BlueAmulet|BlueAmulet]]
+
| [[User:Alsa|Alsa]]
 
| 1x
 
| 1x
|[https://creativecommons.org/licenses/by-nc/4.0/ CC BY-NC 4.0]
+
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 
| ESRGAN
 
| ESRGAN
| Banding
+
| Dithering
| 2020-07-16
+
|  
| N64 textures use a color depth of 5-bits per channel, this model attempts to clean them, restoring smooth gradients in textures
+
| Dithered Images
| 1x_BC1-smooth2.pth
+
| [https://mega.nz/#!7io2gSQR!UB9u2k51daixTgC2H0LdOWzNlkDyIDHwxBX4BVY2J3k JPG (0-20%)]
 
|  
 
|  
 
|-
 
|-
  
 
|-
 
|-
| [https://drive.google.com/drive/folders/1WD2RqtER_jk0QkPFUySeXhDV2aAO8_nr?usp=sharing 1x_Bandage-Smooth]
+
| [https://drive.google.com/file/d/1H4KQyhcknOoExjvDdsoxAgTBMO7JuJ3w/view 4xFSDedither]
| |[[User:Mutin_Choler|Mutin_Choler]]
+
| Jacob
| 1x
+
| 4x
| WTFPL
+
| GNU GPL3
 
| ESRGAN
 
| ESRGAN
| Banding
+
| Dithering
| 2021-04-16
+
| 2020-01-29
| Attempts to remove the damages done by color banding. For this model, I downscaled all of the HR images to make their pixel size closer to 1000x1000 using the Box filter and then downscaled them again by 50% using the Point filter. Afterwards, I applied 64-bit color banding to every image in the dataset.
+
| 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 (https://github.com/ManuelFritsche/real-world-sr/tree/master/esrgan-fs/codes) for better details compared to plain ESRGAN.
| 1xESRGAN
+
| RRDB_ESRGAN_x4
| https://imgsli.com/NTA1NTk/1/0
+
| https://imgsli.com/MTI1NTY/
 
|-
 
|-
  
 
|-
 
|-
| [https://f002.backblazeb2.com/file/ESRGAN/_RELEASE/1x_Debandurh_FS_lite_140000_G.pth Debandurh_FS Ultra-lite (1x_Debandurh_FS_lite_140000_G.pth)]
+
| [https://drive.google.com/open?id=1Vtwt9kgnmju1edibl07WI6Rg7CrAlHIx 4xFSDedither_Manga]
| |[[User:Twittman|Twittman]]
+
| Jacob
| 1x
+
| 4x
| [https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0] (?)
+
| [https://www.gnu.org/licenses/gpl-3.0.en.html GNU GPLv3]
| ESRGAN
+
| ESRGAN-FS
| Debanding images
+
| Dithering
|  
+
| 2020-04-21
| Dataset: Anime, real life, manga
+
| Cartoons/pixel art/other non-realistic stuff with dithering
| None
+
| RRDB_ESRGAN_x4.pth
|  
+
| https://imgsli.com/MTQ3Nzc
 
|-
 
|-
  
|}
+
| [https://buildism.net/files/4xFSDedither_Riven.pth 4xFSDedither_Riven]
 
+
| Jacob
====Halo Removal====
+
| 4x
{| class="wikitable sortable"
+
| [https://www.gnu.org/licenses/gpl-3.0.en.html GNU GPLv3]
 +
| ESRGAN-FS
 +
| Dithering
 +
| 2020-07-11
 +
| Fine-tuned 4xFSDedither 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.
 +
| 4xESRGAN
 +
| https://imgsli.com/MTg5NTM
 +
 
 +
https://imgsli.com/MTg5NTQ
 +
 
 +
https://imgsli.com/MTg5NTU
 +
 
 +
https://i.imgur.com/j7Wtn0G.png
 
|-
 
|-
! Model Name
 
! Author
 
! Scale
 
! License
 
! Architecture
 
! Purpose (short)
 
! Date Posted
 
! Purpose (Full)
 
! Pretrained_Model_G
 
! sample
 
  
 +
| [https://drive.google.com/drive/folders/1XKREcpHnaYz_dcmkMPDbyETcXzDcQsR2?usp=sharing 1x_DitherDeleter-Smooth_104000_G]
 +
| |[[User:Mutin_Choler|Mutin_Choler]]
 +
| 1x
 +
| WTFPL
 +
| ESRGAN
 +
| Dithering
 +
| 2021-01-23
 +
| 1xESRGAN
 +
|
 +
| https://imgsli.com/Mzg0MTk/1/2
 
|-
 
|-
| [https://drive.google.com/drive/u/1/folders/1pT75GdIU4JAH75eJzpxsaHfIoJ1a_gXB 1x_DeEdge_105000_G]
+
 
 +
| [https://drive.google.com/drive/folders/1XKREcpHnaYz_dcmkMPDbyETcXzDcQsR2?usp=sharing 1x_DitherDeleterV3-Smooth]
 
| |[[User:Mutin_Choler|Mutin_Choler]]
 
| |[[User:Mutin_Choler|Mutin_Choler]]
 
| 1x
 
| 1x
 
| WTFPL
 
| WTFPL
 
| ESRGAN
 
| ESRGAN
| Halo remover
+
| Dithering
| 2020-11-03
+
| 2021-04-18
|  
+
| Attempts to remove the damages done by dithering. For this model, I downscaled all of the HR images to make their pixel size closer to 1000x1000 using the Box filter and then downscaled them again by 50% using the Point filter. Afterwards, I applied 32-bit Riemersma to every image in the dataset.
 
| 1xESRGAN
 
| 1xESRGAN
| https://imgsli.com/MjgxMDU/3/0
+
| https://imgsli.com/NTA5MjU/1/0
 
|-
 
|-
 
|}
 
|}
  
====Noise====
+
====Blurring====
  
 
{| class="wikitable sortable"
 
{| class="wikitable sortable"
Line 2,552: Line 2,665:
 
! Purpose (Full)
 
! Purpose (Full)
 
! Pretrained_Model_G
 
! Pretrained_Model_G
! sample
+
! Sample
  
 
|-
 
|-
| [https://drive.google.com/file/d/1GoUY7t5kE0Ubi-NDvlMB8lUs-_u_f6tS/view?usp=sharing LADDIER1]
+
| [https://mega.nz/file/CYoXCAKR#eRKwpP9rCL9vFCsO41oVImvmVwAEOtYUlp2CquX0Fw8 1x_ThePi7on-Solidd_Deborutify_UltraLite_260k_G]
| Alexander Syring
+
| ThePi7on
| 4x
+
| 1x
| Unknown
+
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 
| ESRGAN
 
| ESRGAN
| Denoise
+
| Deblur
| 2019-11-13
+
| 2021-05-09
| Remove noise, grain, box blur, lens blur and gaussian blur and increase overall image quality.
+
| Sharpening, line darkening and slight line thinning, specifically made for the Boruto anime.
| RRDB_ESRGAN_x4.pth
+
|
|  
+
|
 
|-
 
|-
  
| [https://drive.google.com/drive/u/1/folders/12j6BvuR7SU48ui9DHjBPB1vzPpQri9bD 1x_NoiseToner-Poisson-Detailed_108000_G]
+
 
| |[[User:Mutin_Choler|Mutin_Choler]]
+
|-
 +
| [https://mega.nz/file/zcBRHRbJ#iLStv-F5ax-vQBPG7N6OaRv_xBhRHux3pdnSG43GuMg 1x_DeBLR]
 +
| BlackScout
 
| 1x
 
| 1x
| WTFPL
+
| Unknown or None
 
| ESRGAN
 
| ESRGAN
| Noise remover
+
| Deblur
| 2021-10-08
+
| 2020-02-19
| Attempts to remove the damages done from noise. Successor of sorts to Noisetoner_Poisson_150000_G
+
| General Deblurring
| 1xESRGAN
+
|
| https://imgsli.com/NzQ5MzM/2/0
+
|
 
|-
 
|-
  
| [https://drive.google.com/drive/u/1/folders/12j6BvuR7SU48ui9DHjBPB1vzPpQri9bD 1x_NoiseToner_Poisson_150000_G]
+
| [https://cloud.owncube.com/s/p3w7M2sfaWWYCK9/download?path=%2F&files=1x_mdeblur.pth mdeblur]
| |[[User:Mutin_Choler|Mutin_Choler]]
+
| LyonHrt
 
| 1x
 
| 1x
| WTFPL
+
|UNKNOWN
 
| ESRGAN
 
| ESRGAN
| Noise remover
+
| Blurring
| 2020-10-14
+
|
 +
| Strong deblurring model
 +
|  
 
|  
 
|  
| 1xESRGAN
 
| https://imgsli.com/MjU1ODU/3/1
 
 
|-
 
|-
  
| [https://drive.google.com/drive/u/1/folders/12j6BvuR7SU48ui9DHjBPB1vzPpQri9bD 1x_NoiseToner_Uniform_100000_G]
+
| [https://1drv.ms/u/s!Aip-EMByJHY27TDhiFPB39jM0pwq?e=5zAFtO 1x_PixelSharpen_100000]
| |[[User:Mutin_Choler|Mutin_Choler]]
+
| [[User:DinJerr|DinJerr]]
 
| 1x
 
| 1x
| WTFPL
+
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 
| ESRGAN
 
| ESRGAN
| Noise remover
+
| Pixel Art
| 2020-10-14
+
| 2021-03-19
|  
+
| Restores blurry/upscaled pixel art.
 
| 1xESRGAN
 
| 1xESRGAN
| https://imgsli.com/MjU1Nzg/2/1
+
| https://imgsli.com/NDUxNDQ/5/4
 
|-
 
|-
  
|-
+
| [https://1drv.ms/u/s!Aip-EMByJHY28w0Zkd9fdANYH3j6?e=9zwxII 1x_ArtClarity]
| [https://drive.google.com/file/d/1f5kIr3gG5GjbN_EM96CCB0nqu_2_56gh/view 1x_ISO_denoise_v1]
+
| [[User:DinJerr|DinJerr]]
| Alpha
 
 
| 1x
 
| 1x
 
| WTFPL
 
| WTFPL
 
| ESRGAN
 
| ESRGAN
| denoise
+
| Pixel Art
| not on discord
+
| 2021-08-05
| Remove high ISO noise
+
| Texture retaining denoiser and sharpener for digital artwork. Helps resized/generated artwork look more like it is in 'native' resolution.
| 1xESRGAN
+
| 4xPSNR
 +
| [https://1drv.ms/u/s!Aip-EMByJHY28xKFp93VIGydNsoN?e=vN2WYh Sample Gallery]
 +
|-
  
|
+
| [https://de-next.owncube.com/index.php/s/aAojXwLTPZto8rP 1x_Fatality_DeBlur]
|-
+
| |[[User:Twittman|Twittman]]
| [https://drive.google.com/file/d/1HSUvnQm56_CmVAAMjTfVn_Hi6a4Bm2h0/view 1x_ISO_denoise_v2]
 
| Alpha
 
 
| 1x
 
| 1x
| WTFPL
+
| [https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 
| ESRGAN
 
| ESRGAN
| denoise
+
| Deblur
| not on discord
+
| 2019-10-16
| Remove high ISO noise
+
| Dataset: Mix of anime, manga, and photos
| ISO denoise v1
+
| 1x_DeJpeg_Fatality_01_175000_G.pth
 +
|
 +
|-
  
 +
| [https://de-next.owncube.com/index.php/s/nmNfcd33mnLLwEC UnResize V3 (1x_UnResize_V3_200000_G.pth)]
 +
| |[[User:Twittman|Twittman]]
 +
| 1x
 +
| [https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0] (?)
 +
| ESRGAN
 +
| Deblur / Unresize
 +
|
 +
| Dataset: Anime, manga, and photos -
 +
Purpose: Fix images that have been arbitrarily / poorly resized, such as non-integer nearest-neighbor upscaling/downscaling -
 +
Also acts as an image sharpener/deblur when used on slightly soft inputs
 +
| 1x_UnResize_MKII_030000_G.pth
 
|  
 
|  
 
|-
 
|-
| [https://drive.google.com/file/d/1kO3fx9FL5rKhoDk2n4kiiPWuCNHojkIR/view Film-Degrainer_1-000]
+
 
| Tika
+
| [https://mega.nz/folder/OVomUZyT#vDyNntGY1MirBEACNjpqVg 1x-Focus and 1x-Focus_Moderate]
 +
| |[[User:Kim2091|Kim2091]]
 
| 1x
 
| 1x
| CC0
+
| [https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 
| ESRGAN
 
| ESRGAN
| denoise
+
| Deblur
| not on discord
 
| Remove film grain/noise
 
| none
 
 
|  
 
|  
|}
+
| These models deblur most images. It was trained mostly on aniso2 and iso blurring (BSRGAN augmentation) with some gaussian mixed in. It performs well on most blurry images, but I'd recommend using something like Fatality_Deblur for very strong gaussian blur. Dataset: The UniScale Dataset + My Fabric dataset
 +
| 1xESRGAN
 +
| [https://cdn.discordapp.com/attachments/893483065664483338/893483645803847690/unknown.png Sample 1]
 +
 
 +
[https://cdn.discordapp.com/attachments/893483065664483338/893497416207188028/unknown.png Sample 2]
  
 +
[https://cdn.discordapp.com/attachments/893483065664483338/893483280895193118/unknown.png Sample 3]
  
====Oversharpening====
+
[https://cdn.discordapp.com/attachments/893483065664483338/893496207832408135/unknown.png Sample 4]
{| class="wikitable sortable"
 
 
|-
 
|-
! Model Name
 
! Author
 
! Scale
 
! License
 
! Architecture
 
! Purpose (short)
 
! Date Posted
 
! Purpose (Full)
 
! Pretrained_Model_G
 
! sample
 
  
|-
+
| [https://de-next.owncube.com/index.php/s/HyygDgXEB45JqBH ReFocus V3 (1x_ReFocus_V3_140000_G.pth)]
| [https://mega.nz/#!qRZD3SLL!RqP7cISIcoW5YakPya9CXEWSEHWqUSKdWLQSdYKGa14 1x_DeSharpen]
+
| |[[User:Twittman|Twittman]]
| Loinne
+
| 1x
 +
| [https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 +
| ESRGAN
 +
| Deblur / ReFocus
 +
| 2022-03-12
 +
| Dataset: Photos, Anime, and manga    - 
 +
Purpose: DeBlur, ReFocus, Sharpen Real life style images, but will work on Anime images too.
 +
| 1x_Saiyajin_DeJPEG_300000_G.pth
 +
|
 +
|-
 +
 
 +
| [https://de-next.owncube.com/index.php/s/GjbBw7pcgm5gmYm ReFocus Cleanly (1x_ReFocus_Cleanly_100000_G.pth)]
 +
| |[[User:Twittman|Twittman]]
 
| 1x
 
| 1x
|[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]
 
| ESRGAN
 
| ESRGAN
| Denoise
+
| Deblur / ReFocus
| 2019-06-03
+
| 2022-01-21
| 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 oversharpening, 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 oversharpened, for example in image consisting of several combined images, 1 of them with sharpen noise.
+
| Dataset: Anime, manga, and photos   - 
| 1st attempt on random sharpening with the same dataset at 200000 iterations, which was trained on non-random desharp model, total ~600000 iterations on 3 models.
+
Purpose: DeBlur, ReFocus, Sharpen Manga, Anime and cartoon style images, but will work on real life images too.
 +
| 1x_ReFocus_V3_110000_G.pth
 
|  
 
|  
 
|-
 
|-
 +
 
|}
 
|}
  
====DeToon====
+
====Banding====
 
{| class="wikitable sortable"
 
{| class="wikitable sortable"
 
|-
 
|-
Line 2,681: Line 2,814:
 
! Purpose (Full)
 
! Purpose (Full)
 
! Pretrained_Model_G
 
! Pretrained_Model_G
! sample
+
! Sample
 
 
  
 
|-
 
|-
| [https://drive.google.com/open?id=1uJvdx3g3GEY0VxMnHb0ItBvoc6pmGvuH detoon]
+
| [https://drive.google.com/file/d/1GNguVMV4E5t_4dn23w6UMXVKN6Ff9XPt/view?usp=sharing 1x_N64clean]
| |[[User:LyonHrt|LyonHrt]]
+
| |[[User:BlueAmulet|BlueAmulet]]
|  
+
| 1x
|  
+
|[https://creativecommons.org/licenses/by-nc/4.0/ CC BY-NC 4.0]
 
| ESRGAN
 
| ESRGAN
| Detooning
+
| Banding
| 2019-06-24
+
| 2020-07-16
| A toon to realistic shading style model to wiki under drawings
+
| N64 textures use a color depth of 5-bits per channel, this model attempts to clean them, restoring smooth gradients in textures
|  
+
| 1x_BC1-smooth2.pth
 
|  
 
|  
 
|-
 
|-
  
| [https://drive.google.com/file/d/12vkDn35cuVnq2KYpkaA3XfCmk6zY_u4_/view?usp=sharing 1xDoubleDetoon]
+
|-
| Joey
+
| [https://drive.google.com/drive/folders/1WD2RqtER_jk0QkPFUySeXhDV2aAO8_nr?usp=sharing 1x_Bandage-Smooth]
 +
| |[[User:Mutin_Choler|Mutin_Choler]]
 
| 1x
 
| 1x
| CC BY-NC-SA
+
| WTFPL
 
| ESRGAN
 
| ESRGAN
| Detooning
+
| Banding
| 2020-07-31
+
| 2021-04-16
| An attempt to detoon images/drawings of people
+
| Attempts to remove the damages done by color banding. For this model, I downscaled all of the HR images to make their pixel size closer to 1000x1000 using the Box filter and then downscaled them again by 50% using the Point filter. Afterwards, I applied 64-bit color banding to every image in the dataset.
 
| 1xESRGAN
 
| 1xESRGAN
 +
| https://imgsli.com/NTA1NTk/1/0
 +
|-
 +
 +
|-
 +
| [https://f002.backblazeb2.com/file/ESRGAN/_RELEASE/1x_Debandurh_FS_lite_140000_G.pth Debandurh_FS Ultra-lite (1x_Debandurh_FS_lite_140000_G.pth)]
 +
| |[[User:Twittman|Twittman]]
 +
| 1x
 +
| [https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0] (?)
 +
| ESRGAN
 +
| Debanding images
 +
|
 +
| Dataset: Anime, real life, manga
 +
| None
 
|  
 
|  
 
|-
 
|-
 +
 
|}
 
|}
  
 
+
====Halo Removal====
 
 
 
 
====Image De/Colorization====
 
 
{| class="wikitable sortable"
 
{| class="wikitable sortable"
 
|-
 
|-
Line 2,725: Line 2,869:
 
! Purpose (Full)
 
! Purpose (Full)
 
! Pretrained_Model_G
 
! Pretrained_Model_G
! sample
+
! Sample
  
 
|-
 
|-
| [https://drive.google.com/file/d/1UgxnymTfLaVhY3aaDuBRnTZ9ehckKVKJ BS_Colorizer/Vapourizer]
+
| [https://drive.google.com/drive/u/1/folders/1pT75GdIU4JAH75eJzpxsaHfIoJ1a_gXB 1x_DeEdge_105000_G]
| |[[User:BlackScout|BlackScout]]
+
| |[[User:Mutin_Choler|Mutin_Choler]]
 
| 1x
 
| 1x
| [https://www.gnu.org/licenses/gpl-3.0.en.html GNU GPLv3]
+
| WTFPL
 
| ESRGAN
 
| ESRGAN
| Colorization
+
| Halo remover
| 2020-01-30
+
| 2020-11-03
| Colorize Grayscale or Desaturated images. - This model is a partially failed attempt at restoring color from Grayscale | B/W | 100% Desaturated images. It mostly results in Blue and Yellow images with slight hints of Green, Orange and Magenta. You are free to use this as a pretrain to achieve better results.
+
|  
 
| 1xESRGAN
 
| 1xESRGAN
|  
+
| https://imgsli.com/MjgxMDU/3/0
 
|-
 
|-
 +
|}
  
| [https://u.pcloud.link/publink/show?code=kZpAGFXZARxWIu9Pp4YDLUSIN8I8aFuJ5TKV 1x_SpongeColor-Lite]
+
====Noise====
| Joey
 
| 1x
 
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 
| ESRGAN (lite) [nf=32 nb=12]
 
| Colorization
 
| 2020-10-15
 
| The first attempt at ESRGAN colorization that produces more than 2 colors. Doesn't work that great but it was a neat experiment.
 
| none
 
|
 
|-
 
 
 
|-
 
| [https://drive.google.com/open?id=1-mxmDF1Dh-PnQqRz_PeCrvsTkHjYCbi3 Deoldify]
 
| Rastrum
 
| 4x
 
| CC0
 
| ESRGAN
 
| Photos
 
| 2019-08-29
 
| Old black and white photo restoration.
 
| Falcon Fanart
 
|
 
|-
 
|}
 
  
===Stylization===
 
====Images====
 
 
{| class="wikitable sortable"
 
{| class="wikitable sortable"
 
|-
 
|-
Line 2,779: Line 2,898:
 
! Purpose (Full)
 
! Purpose (Full)
 
! Pretrained_Model_G
 
! Pretrained_Model_G
! sample
+
! Sample
  
 
|-
 
|-
| [https://drive.google.com/file/d/1AR2KhmJ23woifnpe1dr0BXkmF7WLG2pS/view?usp=sharing Rek's Effeks Photoanime v2]
+
| [https://drive.google.com/file/d/1GoUY7t5kE0Ubi-NDvlMB8lUs-_u_f6tS/view?usp=sharing LADDIER1]
| Rek
+
| Alexander Syring
 
| 4x
 
| 4x
| GNU GPL3
+
| Unknown
 
| ESRGAN
 
| ESRGAN
| Stylization
+
| Denoise
| 2020-10-17
+
| 2019-11-13
| Photo stylization from JPEGs. Trained on images upscaled by ISO Denoise v2 -> DeJPEG Fatality PlusULTRA -> NMKD Yan2. Essentially a combination of that model chain into one. Use if you're looking for a stylized output, not photo quality.
+
| Remove noise, grain, box blur, lens blur and gaussian blur and increase overall image quality.
| NMKD Yandere2
+
| RRDB_ESRGAN_x4.pth
 
|  
 
|  
 
|-
 
|-
  
 +
| [https://drive.google.com/drive/u/1/folders/12j6BvuR7SU48ui9DHjBPB1vzPpQri9bD 1x_NoiseToner-Poisson-Detailed_108000_G]
 +
| |[[User:Mutin_Choler|Mutin_Choler]]
 +
| 1x
 +
| WTFPL
 +
| ESRGAN
 +
| Noise remover
 +
| 2021-10-08
 +
| Attempts to remove the damages done from noise. Successor of sorts to Noisetoner_Poisson_150000_G
 +
| 1xESRGAN
 +
| https://imgsli.com/NzQ5MzM/2/0
 
|-
 
|-
| [https://drive.google.com/file/d/1URPhkwgraCzWAIkB8Pk4ZPr1LTdfdwxJ/view?usp=sharing Ghibli_Grain.pth]
+
 
| nonogamester#3975
+
| [https://drive.google.com/drive/u/1/folders/12j6BvuR7SU48ui9DHjBPB1vzPpQri9bD 1x_NoiseToner_Poisson_150000_G]
 +
| |[[User:Mutin_Choler|Mutin_Choler]]
 
| 1x
 
| 1x
| [https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
+
| WTFPL
| ESRGAN  
+
| ESRGAN
| Realistic Ghibli Grain
+
| Noise remover
| 2022-11-16
+
| 2020-10-14
| Text or Anime
+
|  
| Attempt to get the nostalgic grain feel of classically animated Ghibli movies. Got the idea from  researching the making of Ronin's :PKT_blue: 1x_AnimeFilmGrain28k. This was made running Nate video denoising with the preset "more denoising" on kiki's delivery service then overlaying it 66.65 percent over the original with DaVinci. On digital drawn anime it gives a slightly [10%] more organic/sharp feel to black lines. Matches well with content that already has a light digital grain. Run Twice for heavy grain. Eternal Thanks to all that indulge all my :psyduck:  Wonderings :element_fire_neon:
+
| 1xESRGAN
| https://slow.pics/c/M6dNTu6G https://imgsli.com/MTM0Nzcw/9/8
+
| https://imgsli.com/MjU1ODU/3/1
 +
|-
  
|}
+
| [https://drive.google.com/drive/u/1/folders/12j6BvuR7SU48ui9DHjBPB1vzPpQri9bD 1x_NoiseToner_Uniform_100000_G]
 
+
| |[[User:Mutin_Choler|Mutin_Choler]]
==Specialized Models==
+
| 1x
===Text===
+
| WTFPL
{| class="wikitable sortable"
 
|-
 
! Model Name
 
! Author
 
! Scale
 
! License
 
! Architecture
 
! Purpose (short)
 
! Date Posted
 
! Purpose (Full)
 
! Pretrained_Model_G
 
! sample
 
 
 
|-
 
| [https://drive.google.com/file/d/15ovbadCoYs7q8nSd5Mq02PqBOpwiBkoS/view 2x_BSTexty]
 
| |[[User:BlackScout|BlackScout]]
 
| 2x
 
| [https://www.gnu.org/licenses/gpl-3.0.en.html GNU GPLv3]
 
 
| ESRGAN
 
| ESRGAN
| Text
+
| Noise remover
| 2020-03-04
+
| 2020-10-14
| As the name might suggest, this model aims to upscale text with less distortion that other models. It seems to do a good job generally, but don't expect it to be a state of the art model that can upscale magazines and stuff. It makes things more readable but since it was train on B/W pictures it desaturates them.
 
| 2x_ESRGAN
 
 
|  
 
|  
 
+
| 1xESRGAN
 +
| https://imgsli.com/MjU1Nzg/2/1
 
|-
 
|-
| [https://mega.nz/folder/rdpkjZzC#eUXPed_vntJKLrB0wpeJ-w 4x-TextSharpV1]
 
| |[[User:Kim2091|Kim2091]]
 
| 4x
 
| [https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 
| ESRGAN
 
| Text
 
| 2021-12-28
 
| Text or Anime
 
| Interpolation between 4x-UltraSharp and 4x-TextSharp-v0.5. This is a duplicate listing of '''4x-AnimeSharp'''
 
| https://media.discordapp.net/attachments/556604553424928768/925526934555881532/unknown.png https://cdn.discordapp.com/attachments/903415274521374750/925533775616696340/unknown.png
 
 
  
 
|-
 
|-
| [https://icedrive.net/1/43GNBihZyi NMKD Typescale]
+
| [https://drive.google.com/file/d/1f5kIr3gG5GjbN_EM96CCB0nqu_2_56gh/view 1x_ISO_denoise_v1]
| |[[User:nmkd|Nmkd]]
+
| Alpha
| 8x
+
| 1x
 
| WTFPL
 
| WTFPL
 
| ESRGAN
 
| ESRGAN
| Text
+
| denoise
| 2020-11-04
+
| not on discord
| Low-resolution text/typography and symbols
+
| Remove high ISO noise
| 8xESRGAN
+
| 1xESRGAN
 +
 
 +
|
 +
|-
 +
| [https://drive.google.com/file/d/1HSUvnQm56_CmVAAMjTfVn_Hi6a4Bm2h0/view 1x_ISO_denoise_v2]
 +
| Alpha
 +
| 1x
 +
| WTFPL
 +
| ESRGAN
 +
| denoise
 +
| not on discord
 +
| Remove high ISO noise
 +
| ISO denoise v1
 +
 
 +
|
 +
|-
 +
| [https://drive.google.com/file/d/1kO3fx9FL5rKhoDk2n4kiiPWuCNHojkIR/view Film-Degrainer_1-000]
 +
| Tika
 +
| 1x
 +
| CC0
 +
| ESRGAN
 +
| denoise
 +
| not on discord
 +
| Remove film grain/noise
 +
| none
 +
|
 +
 
 
|-
 
|-
 +
| [https://1drv.ms/u/s!Aip-EMByJHY2gYQUcbSTFgrdwtMjQA?e=A5p6lH 1x_Fatality_NoiseToner]
 +
| [https://upscale.wiki/wiki/User:DinJerr DinJerr]
 +
| 1x
 +
| WTFPL
 +
| ESRGAN
 +
| sharpen & denoise
 +
| 2023-04-16
 +
| Interpolation of Mutin_Choler's various 1x_NoiseToner_Poisson with Twittman's 1x_Fatality_DeBlur
 +
| none
 +
|
 
|}
 
|}
  
===Inpainting===
+
====Oversharpening====
 
{| class="wikitable sortable"
 
{| class="wikitable sortable"
 
|-
 
|-
Line 2,873: Line 3,010:
 
! Purpose (Full)
 
! Purpose (Full)
 
! Pretrained_Model_G
 
! Pretrained_Model_G
! sample
+
! Sample
  
 
|-
 
|-
| [https://e.pcloud.link/publink/show?code=kZQOu7ZldzmFyMPUcFNGkEvwqOxQ8Bl3CeX 1x_sudo_inpaint_PartialConv2D_424000_G.pth]
+
| [https://mega.nz/#!qRZD3SLL!RqP7cISIcoW5YakPya9CXEWSEHWqUSKdWLQSdYKGa14 1x_DeSharpen]
| sudo rm -rf / --no-preserve-root#8353
+
| Loinne
 
| 1x
 
| 1x
| [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]
 
| ESRGAN
 
| ESRGAN
| Inpainting
+
| Denoise
| 2020-12-05
+
| 2019-06-03
| Experimental PartialConv2D attempt to paint with ESRGAN. Took ~10.4 days of training on a P100 and around 1.5 months in total due to Colab limits. Not sure if I will continue training it since training is very slow, but may get better.. Warning: Result can vary with different tilesizes. Try not to tile your data.
+
| 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 oversharpening, 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 oversharpened, for example in image consisting of several combined images, 1 of them with sharpen noise.
| None
+
| 1st attempt on random sharpening with the same dataset at 200000 iterations, which was trained on non-random desharp model, total ~600000 iterations on 3 models.
| [https://e.pcloud.link/publink/show?code=kZQOu7ZldzmFyMPUcFNGkEvwqOxQ8Bl3CeX Samples]
 
 
 
|-
 
| [https://icedrive.net/1/43GNBihZyi 1x_NMKD-YandereInpaint_375000_G.pth]
 
| [[User:nmkd|Nmkd]]
 
| 1x
 
| Unknown
 
| ESRGAN
 
| Inpainting
 
| Unknown
 
 
|  
 
|  
| None
 
| [https://icedrive.net/1/43GNBihZyi Samples]
 
 
|-
 
|-
 
 
|}
 
|}
  
===Fabric/Cloth===
+
====DeToon====
 
{| class="wikitable sortable"
 
{| class="wikitable sortable"
 
|-
 
|-
Line 2,914: Line 3,038:
 
! Purpose (Full)
 
! Purpose (Full)
 
! Pretrained_Model_G
 
! Pretrained_Model_G
! sample
+
! Sample
 +
 
  
 
|-
 
|-
| [https://mega.nz/folder/6EJjwADL#oDvv9p0z1t9nQvObbvBRBQ 4x-Fabric and Fabric-Alt]
+
| [https://drive.google.com/open?id=1uJvdx3g3GEY0VxMnHb0ItBvoc6pmGvuH detoon]
| |[[User:Kim2091|Kim2091]]
+
| |[[User:LyonHrt|LyonHrt]]
| 4x
+
|  
| [https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
+
|  
 
| ESRGAN
 
| ESRGAN
| Fabric
+
| Detooning
| 2021-09-15
+
| 2019-06-24
| This model set upscales fabric or cloth textures (works on cats too!). The Alt model is just an earlier iteration version. It may work better on some images.The images need to be minimally compressed or passed through a decompression model first. It works with DDS compression though.
+
| A toon to realistic shading style model to wiki under drawings
| 4xESRGAN
+
|
| [https://cdn.discordapp.com/attachments/887850596214902835/887852703873638410/unknown.png Sample 1] [https://cdn.discordapp.com/attachments/887850596214902835/887852170156847124/unknown.png Sample 2] [https://cdn.discordapp.com/attachments/887850596214902835/887856392269094952/unknown.png Sample 3] [https://cdn.discordapp.com/attachments/887850596214902835/887853344545194084/unknown.png Sample 4]
+
|
 
+
|-
|}
+
 
 +
| [https://drive.google.com/file/d/12vkDn35cuVnq2KYpkaA3XfCmk6zY_u4_/view?usp=sharing 1xDoubleDetoon]
 +
| Joey
 +
| 1x
 +
| CC BY-NC-SA
 +
| ESRGAN
 +
| Detooning
 +
| 2020-07-31
 +
| An attempt to detoon images/drawings of people
 +
| 1xESRGAN
 +
|
 +
|-
 +
|}
 +
 
  
  
===Alphas===
+
 
 +
====Image De/Colorization====
 
{| class="wikitable sortable"
 
{| class="wikitable sortable"
 
|-
 
|-
Line 2,943: Line 3,082:
 
! Purpose (Full)
 
! Purpose (Full)
 
! Pretrained_Model_G
 
! Pretrained_Model_G
! sample
+
! Sample
  
 
|-
 
|-
|[https://drive.google.com/open?id=192Mw2_yUwCgqt3tAJ2sRxZ8hYHe4CKrZ FireAlpha]
+
| [https://drive.google.com/file/d/1UgxnymTfLaVhY3aaDuBRnTZ9ehckKVKJ BS_Colorizer/Vapourizer]
| |[[User:BlueAmulet|BlueAmulet]]
+
| |[[User:BlackScout|BlackScout]]
| 4x
+
| 1x
|[https://creativecommons.org/licenses/by-nc/4.0/ CC BY-NC 4.0]
+
| [https://www.gnu.org/licenses/gpl-3.0.en.html GNU GPLv3]
 
| ESRGAN
 
| ESRGAN
| Alpha (4 channel)
+
| Colorization
| 2019-11-12
+
| 2020-01-30
|  
+
| Colorize Grayscale or Desaturated images. - This model is a partially failed attempt at restoring color from Grayscale | B/W | 100% Desaturated images. It mostly results in Blue and Yellow images with slight hints of Green, Orange and Magenta. You are free to use this as a pretrain to achieve better results.
| none
+
| 1xESRGAN
 
|  
 
|  
 
|-
 
|-
  
| [https://u.pcloud.link/publink/show?code=kZQfN4XZmIrkODabBWhtmmA95c0GMb19WUuy 4x_1ch-Alpha-Lite]
+
| [https://u.pcloud.link/publink/show?code=kZpAGFXZARxWIu9Pp4YDLUSIN8I8aFuJ5TKV 1x_SpongeColor-Lite]
 
| Joey
 
| Joey
| 4x
+
| 1x
 
|[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]
| ESRGAN
+
| ESRGAN (lite) [nf=32 nb=12]
| Alpha
+
| Colorization
| 2020-12-15
+
| 2020-10-15
| Obsoleted by Joey's Fork - Alpha channels of PNGs
+
| The first attempt at ESRGAN colorization that produces more than 2 colors. Doesn't work that great but it was a neat experiment.
 
| none
 
| none
 
|  
 
|  
 
|-
 
|-
  
| [https://u.pcloud.link/publink/show?code=kZkR7hXZhfCA4WvVQvbxXkNN50tazuxj73uX 2x_Gen5-Alpha]
+
|-
| Joey
+
| [https://drive.google.com/open?id=1-mxmDF1Dh-PnQqRz_PeCrvsTkHjYCbi3 Deoldify]
| 2x
+
| Rastrum
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
+
| 4x
 +
| CC0
 
| ESRGAN
 
| ESRGAN
| Pixel Art with Tranparency / Alpha Channel
+
| Photos
| 2021-02-03
+
| 2019-08-29
|  
+
| Old black and white photo restoration.
| 4xFireAlpha
+
| Falcon Fanart
 
|  
 
|  
 
|-
 
|-
 +
|}
  
| [https://u.pcloud.link/publink/show?code=kZeNFhXZbuw2jGJOGqQnKecb8Moo1ha47W57 8x_Sphax-Alpha-NN]
+
===Stylization===
| Joey
+
====Images====
| 8x
+
{| class="wikitable sortable"
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 
| ESRGAN
 
| Pixel Art with Tranparency / Alpha Channel
 
| 2021-02-04
 
| Replica of sphax with transparency
 
| 4xFireAlpha
 
|  
 
 
|-
 
|-
 +
! Model Name
 +
! Author
 +
! Scale
 +
! License
 +
! Architecture
 +
! Purpose (short)
 +
! Date Posted
 +
! Purpose (Full)
 +
! Pretrained_Model_G
 +
! Sample
  
| [https://u.pcloud.link/publink/show?code=kZrIFhXZxnpi72FuwgLz7FIDw8kdrp2suIEk 4x_PocketMonsters-Alpha]
+
|-
| Joey
+
| [https://drive.google.com/file/d/1AR2KhmJ23woifnpe1dr0BXkmF7WLG2pS/view?usp=sharing Rek's Effeks Photoanime v2]
 +
| Rek
 
| 4x
 
| 4x
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
+
| GNU GPL3
 
| ESRGAN
 
| ESRGAN
| Pixel Art with Tranparency / Alpha Channel
+
| Stylization
| 2021-02-04
+
| 2020-10-17
| upscaling pixel art with alpha channels that should perform better than any other. It should work well on both cartoon and 3D styled content.
+
| Photo stylization from JPEGs. Trained on images upscaled by ISO Denoise v2 -> DeJPEG Fatality PlusULTRA -> NMKD Yan2. Essentially a combination of that model chain into one. Use if you're looking for a stylized output, not photo quality.
| 4xFireAlpha
+
| NMKD Yandere2
 +
|
 +
|-
 +
 
 +
|-
 +
| [https://drive.google.com/file/d/1URPhkwgraCzWAIkB8Pk4ZPr1LTdfdwxJ/view?usp=sharing Ghibli_Grain.pth]
 +
| nonogamester#3975
 +
| 1x
 +
| [https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 +
| ESRGAN
 +
| Realistic Ghibli Grain
 +
| 2022-11-16
 +
| Text or Anime. Attempt to get the nostalgic grain feel of classically animated Ghibli movies. Got the idea from  researching the making of Ronin's :PKT_blue: 1x_AnimeFilmGrain28k. This was made running Nate video denoising with the preset "more denoising" on kiki's delivery service then overlaying it 66.65 percent over the original with DaVinci. On digital drawn anime it gives a slightly [10%] more organic/sharp feel to black lines. Matches well with content that already has a light digital grain. Run Twice for heavy grain. Eternal Thanks to all that indulge all my :psyduck:  Wonderings :element_fire_neon:
 
|  
 
|  
 +
| https://slow.pics/c/M6dNTu6G
 +
 +
https://imgsli.com/MTM0Nzcw/9/8
 
|-
 
|-
  
| [https://drive.google.com/file/d/1trYs4AuC9s2JWAbryHdNyy5cgYV-V8cH/view Skyrim Alpha]
+
|-
| Deorder
+
| [https://1drv.ms/u/s!Aip-EMByJHY282MGb4GvQh-BO-SH?e=Hm49eo 1x_ReDetail_v2_126000_G]
| 4x
+
| [[User:DinJerr|DinJerr]]
|[https://creativecommons.org/publicdomain/zero/1.0/ CC0]
+
| 1x
 +
| WTFPL
 
| ESRGAN
 
| ESRGAN
| Pixel Art with Tranparency / Alpha Channel
+
| Digital Illustration
 +
| 2021-02-11
 +
| A failed model that's supposed to insert details into paintings. It's actual use is to interpolate with other 1x models such as with 1x_DoubleDetoon to reduce its colour warping, or with 1x_ArtClarity to emphasize more on feature extraction.
 +
| 1x_ArtClarity
 
|  
 
|  
| Dataset: Alpha Channels from Skyrim
+
 
| Unknown
 
|
 
|-
 
 
|}
 
|}
  
===CGI===
+
==Specialized Models==
 +
===Text===
 
{| class="wikitable sortable"
 
{| class="wikitable sortable"
 
|-
 
|-
Line 3,031: Line 3,193:
 
! Purpose (Full)
 
! Purpose (Full)
 
! Pretrained_Model_G
 
! Pretrained_Model_G
! sample
+
! Sample
 +
 
 
|-
 
|-
 
+
| [https://drive.google.com/file/d/15ovbadCoYs7q8nSd5Mq02PqBOpwiBkoS/view 2x_BSTexty]
| [https://drive.google.com/drive/folders/1m7fO9ywNNRc2FZBKycJiC556rBJpbGoH 1xRedImage10000.pth]
+
| |[[User:BlackScout|BlackScout]]
| 3majsie1995
+
| 2x
| 1x
+
| [https://www.gnu.org/licenses/gpl-3.0.en.html GNU GPLv3]
|[https://creativecommons.org/licenses/by-nc/4.0/ CC BY-NC 4.0]
 
 
| ESRGAN
 
| ESRGAN
| Correct old color photos that are tinted red
+
| Text
| 2022-08-04
+
| 2020-03-04
| Correct old color photos that are tinted red. Model only tested in nature.
+
| As the name might suggest, this model aims to upscale text with less distortion that other models. It seems to do a good job generally, but don't expect it to be a state of the art model that can upscale magazines and stuff. It makes things more readable but since it was train on B/W pictures it desaturates them.
| 1xESRGAN
+
| 2x_ESRGAN
| https://i.imgur.com/5O74rRL.png
+
|  
 +
 
 
|-
 
|-
 
+
| [https://mega.nz/folder/rdpkjZzC#eUXPed_vntJKLrB0wpeJ-w 4x-TextSharpV1]
| [https://1drv.ms/u/s!Aip-EMByJHY20wpGoLuSRzjdqh0T WaifuGAN_v3_30000]
+
| |[[User:Kim2091|Kim2091]]
| DinnJerr
 
 
| 4x
 
| 4x
|[https://creativecommons.org/licenses/by-nc/4.0/ CC BY-NC 4.0]
+
| [https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 
| ESRGAN
 
| ESRGAN
| CGI
+
| Text
| 2019-07-12
+
| 2021-12-28
| Upscaling CG-painted anime with variable outlines.
+
| Text or Anime
| Manga109v2.pth
+
| Interpolation between 4x-UltraSharp and 4x-TextSharp-v0.5. This is a duplicate listing of '''4x-AnimeSharp'''
|
+
| [https://media.discordapp.net/attachments/556604553424928768/925526934555881532/unknown.png Sample 1]
|-
+
 
 +
[https://cdn.discordapp.com/attachments/903415274521374750/925533775616696340/unknown.png Sample 2]
  
  
| [https://drive.google.com/file/d/1zlc8BWOcpCbSZpwf7ArMCqpb0Tf9DlZk/view?usp=drivesdk 2xBS_Wolly]
 
| |[[User:BlackScout|BlackScout]]
 
| 2x
 
| [https://www.gnu.org/licenses/gpl-3.0.en.html GNU GPLv3]
 
| ESRGAN
 
| CGI Animation
 
| 2020-02-07
 
| Pixar Movies or Wall-E pictures/frames
 
| 2xESRGAN
 
|
 
 
|-
 
|-
 
+
| [https://icedrive.net/1/43GNBihZyi NMKD Typescale]
| [https://drive.google.com/file/d/1kNSz5f2_krdlsJ700DwH78OEPYkkcjW5/view?usp=sharing 4xSGI]
+
| |[[User:nmkd|Nmkd]]
| ChrisNonyminus
+
| 8x
| 4x
+
| WTFPL
| GNU GPL3
 
 
| ESRGAN
 
| ESRGAN
| CGI
+
| Text
| 2020-12-17
+
| 2020-11-04
| Upscaling and dedithering pre-rendered sprites, images and textures made in the 90s. Basically vintage CGI.
+
| Low-resolution text/typography and symbols
| 4xESRGAN
+
| 8xESRGAN
| https://imgur.com/a/iwLTjeB
 
|-
 
 
 
| [https://lbry.tv/@Madiator2011:e/x2_CGIMaster_v1:6 CGIMaster_v1]
 
| Madiator
 
| 2x
 
| GNU GPL3
 
| ESRGAN
 
| CGI Animation
 
| 2021-01-14
 
| mixed 3D/2D CGI animations on some 2D animations it might sharpen the edges. General usage is to upscale CGI animations compressed by YouTube.
 
| Sharp_Anime_v2
 
| https://imgsli.com/Mzc1NTM
 
 
|-
 
|-
 
|}
 
|}
  
===Luminance/Chroma===
+
===Inpainting===
 
{| class="wikitable sortable"
 
{| class="wikitable sortable"
 
|-
 
|-
Line 3,108: Line 3,247:
 
! Purpose (Full)
 
! Purpose (Full)
 
! Pretrained_Model_G
 
! Pretrained_Model_G
! sample
+
! Sample
 
 
  
 
|-
 
|-
| [https://drive.google.com/file/d/1FDb9H6GOmE3aRvZbx_-kTiCLS77WyjgL/view 1x_BSLuma]
+
| [https://e.pcloud.link/publink/show?code=kZQOu7ZldzmFyMPUcFNGkEvwqOxQ8Bl3CeX 1x_sudo_inpaint_PartialConv2D_424000_G.pth]
| |[[User:BlackScout|BlackScout]]
+
| sudo rm -rf / --no-preserve-root#8353
 
| 1x
 
| 1x
| [https://www.gnu.org/licenses/gpl-3.0.en.html GNU GPLv3]
+
| [https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 
| ESRGAN
 
| ESRGAN
| Luma
+
| Inpainting
| 2020-02-28
+
| 2020-12-05
| Fix Luminance issues?? LumaSharpen ESRGAN Edition? - This model mostly does what "Lumasharpen" algorithms do. It may help fixing images with Luminance images issues? Like old DVD rips? I am not sure, didn't test.
+
| Experimental PartialConv2D attempt to paint with ESRGAN. Took ~10.4 days of training on a P100 and around 1.5 months in total due to Colab limits. Not sure if I will continue training it since training is very slow, but may get better.. Warning: Result can vary with different tilesizes. Try not to tile your data.
| none
+
| None
|  
+
| [https://e.pcloud.link/publink/show?code=kZQOu7ZldzmFyMPUcFNGkEvwqOxQ8Bl3CeX Samples]
 +
 
 
|-
 
|-
 
+
| [https://icedrive.net/1/43GNBihZyi 1x_NMKD-YandereInpaint_375000_G.pth]
| [https://drive.google.com/file/d/12v_2OwrB_rr6_0tMshHaHgD_XWRNgIyF/view 1x_BSChroma]
+
| [[User:nmkd|Nmkd]]
| |[[User:BlackScout|BlackScout]]
 
 
| 1x
 
| 1x
| [https://www.gnu.org/licenses/gpl-3.0.en.html GNU GPLv3]
+
| Unknown
 
| ESRGAN
 
| ESRGAN
| Sharpen
+
| Inpainting
| 2020-02-29
+
| Unknown
| ChromaSharpen - makes the colors slightly more vibrant with a sideeffect of possibly adding Chromatic Abberation to the image. I am not sure about the usage of this model on real case scenarios but anything blurry or with fuzzy colors could work
 
| none
 
 
|  
 
|  
 +
| None
 +
| [https://icedrive.net/1/43GNBihZyi Samples]
 
|-
 
|-
 +
 
|}
 
|}
  
===Cats===
+
===Fabric/Cloth===
 
{| class="wikitable sortable"
 
{| class="wikitable sortable"
 
|-
 
|-
Line 3,149: Line 3,288:
 
! Purpose (Full)
 
! Purpose (Full)
 
! Pretrained_Model_G
 
! Pretrained_Model_G
! sample
+
! Sample
  
 
|-
 
|-
| |[https://f002.backblazeb2.com/file/ESRGAN/_RELEASE/4x_cat_patch_325000_G.pth Cat_Patch]
+
| [https://mega.nz/folder/6EJjwADL#oDvv9p0z1t9nQvObbvBRBQ 4x-Fabric and Fabric-Alt]
| |[[User:Twittman|Twittman]]
+
| |[[User:Kim2091|Kim2091]]
 
| 4x
 
| 4x
|[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]
 
| ESRGAN
 
| ESRGAN
| Cats
+
| Fabric
| not on discord
+
| 2021-09-15
|  
+
| This model set upscales fabric or cloth textures (works on cats too!). The Alt model is just an earlier iteration version. It may work better on some images.The images need to be minimally compressed or passed through a decompression model first. It works with DDS compression though.
| previous attempt
+
| 4xESRGAN
|
+
| [https://cdn.discordapp.com/attachments/887850596214902835/887852703873638410/unknown.png Sample 1]
|-
+
 
 +
[https://cdn.discordapp.com/attachments/887850596214902835/887852170156847124/unknown.png Sample 2]
 +
 
 +
[https://cdn.discordapp.com/attachments/887850596214902835/887856392269094952/unknown.png Sample 3]
 +
 
 +
[https://cdn.discordapp.com/attachments/887850596214902835/887853344545194084/unknown.png Sample 4]
 +
 
 
|}
 
|}
  
===Coins===
+
 
 +
===Alphas===
 
{| class="wikitable sortable"
 
{| class="wikitable sortable"
 
|-
 
|-
Line 3,177: Line 3,323:
 
! Purpose (Full)
 
! Purpose (Full)
 
! Pretrained_Model_G
 
! Pretrained_Model_G
! sample
+
! Sample
  
 
|-
 
|-
| [https://drive.google.com/file/d/1zC97RLUg5ejqb5-4Dj5q-SzdLPLiMozH/view 4x_Nickelfront]
+
|[https://drive.google.com/open?id=192Mw2_yUwCgqt3tAJ2sRxZ8hYHe4CKrZ FireAlpha]
| BlackScout
+
| |[[User:BlueAmulet|BlueAmulet]]
 
| 4x
 
| 4x
| GNU GPLv3
+
|[https://creativecommons.org/licenses/by-nc/4.0/ CC BY-NC 4.0]
 
| ESRGAN
 
| ESRGAN
| Coins
+
| Alpha (4 channel)
| 2020-02-13
+
| 2019-11-12
| Upscale coins. That's it. If you were mad at me because Nickelback doesn't make any sense. Now you have the perfect solution to your problems. If you want to upscale nickels or anything with similar texture made out of metal, now you can. It works pretty well for a joke.
+
|  
| 4xESRGAN
+
| none
 
|  
 
|  
 
|-
 
|-
|}
 
  
===Faces===
+
| [https://u.pcloud.link/publink/show?code=kZQfN4XZmIrkODabBWhtmmA95c0GMb19WUuy 4x_1ch-Alpha-Lite]
{| class="wikitable sortable"
+
| Joey
 +
| 4x
 +
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 +
| ESRGAN
 +
| Alpha
 +
| 2020-12-15
 +
| Obsoleted by Joey's Fork - Alpha channels of PNGs
 +
| none
 +
|  
 
|-
 
|-
! Model Name
 
! Author
 
! Scale
 
! License
 
! Architecture
 
! Purpose (short)
 
! Date Posted
 
! Purpose (Full)
 
! Pretrained_Model_G
 
! sample
 
  
|-
+
| [https://u.pcloud.link/publink/show?code=kZkR7hXZhfCA4WvVQvbxXkNN50tazuxj73uX 2x_Gen5-Alpha]
| [https://de-next.owncube.com/index.php/s/PF5GTJ9jix4Log8 4x_Faces_N_250000.pth]
+
| Joey
| |[[User:Twittman|Twittman]]
+
| 2x
| 4x
+
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
|
+
| ESRGAN
| ESRGAN
+
| Pixel Art with Tranparency / Alpha Channel
| Faces
+
| 2021-02-03
| 2019-07-18
 
| Upscales images of faces of different scales, sometimes resulting in monsters.
 
| 4xESRGAN
 
 
|  
 
|  
|-
+
| 4xFireAlpha
 
 
|-
 
| [https://de-next.owncube.com/index.php/s/PF5GTJ9jix4Log8 4x_Faces_04_N_180000_G]
 
| |[[User:Twittman|Twittman]]
 
| 4x
 
|
 
| ESRGAN
 
| Faces
 
| 2019-07-23
 
| Upscale faces both pixelized and real
 
| 4x_Faces_N_250000.pth
 
 
|  
 
|  
 
|-
 
|-
  
|-
+
| [https://u.pcloud.link/publink/show?code=kZeNFhXZbuw2jGJOGqQnKecb8Moo1ha47W57 8x_Sphax-Alpha-NN]
| [https://de-next.owncube.com/index.php/s/mZRG4HB3KdP2iP6 Face-Ality V1 (4x_Fatality_Faces_310000_G.pth)]
+
| Joey
| |[[User:Twittman|Twittman]]
+
| 8x
| 4x
+
|[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] (?)
 
 
| ESRGAN
 
| ESRGAN
| Faces
+
| Pixel Art with Tranparency / Alpha Channel
|  
+
| 2021-02-04
| Dataset: Custom (Faces)
+
| Replica of sphax with transparency
| 4x_Faces_04_N_180000_G.pth
+
| 4xFireAlpha
 
|  
 
|  
 
|-
 
|-
  
|-
+
| [https://u.pcloud.link/publink/show?code=kZrIFhXZxnpi72FuwgLz7FIDw8kdrp2suIEk 4x_PocketMonsters-Alpha]
| [https://1drv.ms/u/s!Aip-EMByJHY29BRdcYSnTpgx0LWd?e=I6x5R7 4x_SmolFace_200k]
+
| Joey
| DinJerr
 
 
| 4x
 
| 4x
 
|[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]
 
| ESRGAN
 
| ESRGAN
| Art/People
+
| Pixel Art with Tranparency / Alpha Channel
| 2022-08-19
+
| 2021-02-04
| A sharp upscaler trained specifically for small sprite faces. Does not blend, so avoid using on painted/photo portraits unless you were trying to retain more of the outlines somehow.
+
| upscaling pixel art with alpha channels that should perform better than any other. It should work well on both cartoon and 3D styled content.
| 4x_NMKD-UltraYandere_300k
+
| 4xFireAlpha
| https://imgsli.com/MTIxNjU1/8/9
+
|  
 
|-
 
|-
  
|-
+
| [https://drive.google.com/file/d/1trYs4AuC9s2JWAbryHdNyy5cgYV-V8cH/view Skyrim Alpha]
| [https://1drv.ms/u/s!Aip-EMByJHY27CHg3-1Uue5KgbtK?e=6k0COU BigFace_v3, BigFace_V3_Blend, BigFace_V3_Clear]
+
| Deorder
| DinJerr
 
 
| 4x
 
| 4x
|[https://creativecommons.org/licenses/by-nc/4.0/ CC BY-NC 4.0]
+
|[https://creativecommons.org/publicdomain/zero/1.0/ CC0]
 
| ESRGAN
 
| ESRGAN
| Art/People
+
| Pixel Art with Tranparency / Alpha Channel
| 2020-07-22
+
|  
 +
| Dataset: Alpha Channels from Skyrim
 +
| Unknown
 
|  
 
|  
| 4x_BigFace
 
|
 
 
|-
 
|-
 +
|}
  
 +
===CGI===
 +
{| class="wikitable sortable"
 
|-
 
|-
| [https://drive.google.com/open?id=1OyOJIW224hBhb-aTCbuUQb0qzKmE4oH6 TGHQFace8x]
+
! Model Name
| tg
+
! Author
| 8x
+
! Scale
| GNU GPL3
+
! License
| ESRGAN
+
! Architecture
| Faces
+
! Purpose (short)
| 2019-11-27
+
! Date Posted
| Upscales blurry 128px faces, usefull for enhancing that someone in a small picture.
+
! Purpose (Full)
| 8xESRGAN
+
! Pretrained_Model_G
| https://imgsli.com/OTM1NQ
+
! Sample
 
|-
 
|-
  
| [https://onedrive.live.com/?authkey=%21AP69c97OwKxR3lM&id=36762472C0107E2A%2110699&cid=36762472C0107E2A FArtFace]
+
| [https://drive.google.com/drive/folders/1m7fO9ywNNRc2FZBKycJiC556rBJpbGoH 1xRedImage10000.pth]
| |[[User:DinJerr|DinJerr]]
+
| 3majsie1995
 +
| 1x
 +
|[https://creativecommons.org/licenses/by-nc/4.0/ CC BY-NC 4.0]
 +
| ESRGAN
 +
| Correct old color photos that are tinted red
 +
| 2022-08-04
 +
| Correct old color photos that are tinted red. Model only tested in nature.
 +
| 1xESRGAN
 +
| https://i.imgur.com/5O74rRL.png
 +
|-
 +
 
 +
| [https://1drv.ms/u/s!Aip-EMByJHY20wpGoLuSRzjdqh0T WaifuGAN_v3_30000]
 +
| [[User:DinJerr|DinJerr]]
 
| 4x
 
| 4x
 
|[https://creativecommons.org/licenses/by-nc/4.0/ CC BY-NC 4.0]
 
|[https://creativecommons.org/licenses/by-nc/4.0/ CC BY-NC 4.0]
 
| ESRGAN
 
| ESRGAN
| Faces
+
| CGI
| 2019-11-28
+
| 2019-07-12
|  
+
| Upscaling CG-painted anime with variable outlines.
|  
+
| Manga109v2.pth
 
|  
 
|  
 
|-
 
|-
  
| [https://drive.google.com/file/d/14daAyYuoX-nZ_Lo1U6oekbFqrZTQlQkG/view?usp=drivesdk 4x_BS_SbeveHarvey]
+
 
 +
| [https://drive.google.com/file/d/1zlc8BWOcpCbSZpwf7ArMCqpb0Tf9DlZk/view?usp=drivesdk 2xBS_Wolly]
 
| |[[User:BlackScout|BlackScout]]
 
| |[[User:BlackScout|BlackScout]]
| 4x
+
| 2x
 
| [https://www.gnu.org/licenses/gpl-3.0.en.html GNU GPLv3]
 
| [https://www.gnu.org/licenses/gpl-3.0.en.html GNU GPLv3]
 
| ESRGAN
 
| ESRGAN
| Faces
+
| CGI Animation
| 2020-07-24
+
| 2020-02-07
| Upscale Steve Harvey, but maybe other things, somehow??
+
| Pixar Movies or Wall-E pictures/frames
| 4xESRGAN
+
| 2xESRGAN
 
|  
 
|  
 
|-
 
|-
  
|-
+
| [https://drive.google.com/file/d/1kNSz5f2_krdlsJ700DwH78OEPYkkcjW5/view?usp=sharing 4xSGI]
| [https://drive.google.com/file/d/19ICMKNuS4PbhmtA7Be9lUE6v8NHDDEzP/view Face Focus]
+
| ChrisNonyminus
| |[[User:LyonHrt|LyonHrt]]
 
 
| 4x
 
| 4x
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
+
| GNU GPL3
 
| ESRGAN
 
| ESRGAN
| Faces
+
| CGI
| not on discord
+
| 2020-12-17
| Face De-blur - slightly out of focus / blurred images of faces. It is aimed at faces / hair
+
| Upscaling and dedithering pre-rendered sprites, images and textures made in the 90s. Basically vintage CGI.
| 4xPSNR
+
| 4xESRGAN
|  
+
| https://imgur.com/a/iwLTjeB
 
|-
 
|-
  
|-
+
| [https://lbry.tv/@Madiator2011:e/x2_CGIMaster_v1:6 CGIMaster_v1]
| [https://drive.google.com/file/d/1OyOJIW224hBhb-aTCbuUQb0qzKmE4oH6/view TGHQFace8x]
+
| Madiator
| |[[User:Torrentguy|Torrentguy]]
+
| 2x
| 8x
+
| GNU GPL3
| [https://www.gnu.org/licenses/gpl-3.0.en.html [https://www.gnu.org/licenses/gpl-3.0.en.html GNU GPLv3]]
 
 
| ESRGAN
 
| ESRGAN
| Face Upscaling
+
| CGI Animation
|  
+
| 2021-01-14
|  
+
| mixed 3D/2D CGI animations on some 2D animations it might sharpen the edges. General usage is to upscale CGI animations compressed by YouTube.
| 8xESRGAN
+
| Sharp_Anime_v2
|  
+
| https://imgsli.com/Mzc1NTM
 
|-
 
|-
 
|}
 
|}
  
 
+
===Luminance/Chroma===
===Skin===
 
 
{| class="wikitable sortable"
 
{| class="wikitable sortable"
 
|-
 
|-
Line 3,349: Line 3,488:
 
! Purpose (Full)
 
! Purpose (Full)
 
! Pretrained_Model_G
 
! Pretrained_Model_G
! sample
+
! Sample
  
  
 
|-
 
|-
| [https://drive.google.com/drive/folders/1VkT6tpbCPn2gKZYPtawDJGMpLg6EyRpO?usp=sharing x1_ITF_SkinDiffDetail_Lite_v1]
+
| [https://drive.google.com/file/d/1FDb9H6GOmE3aRvZbx_-kTiCLS77WyjgL/view 1x_BSLuma]
| intheflesh#3116
+
| |[[User:BlackScout|BlackScout]]
 
| 1x
 
| 1x
| CC BY-NC-SA 4.0
+
| [https://www.gnu.org/licenses/gpl-3.0.en.html GNU GPLv3]
 
| ESRGAN
 
| ESRGAN
| Skin Upscaling
+
| Luma
 +
| 2020-02-28
 +
| Fix Luminance issues?? LumaSharpen ESRGAN Edition? - This model mostly does what "Lumasharpen" algorithms do. It may help fixing images with Luminance images issues? Like old DVD rips? I am not sure, didn't test.
 +
| none
 
|  
 
|  
| Adds plausible high frequency detail and removes subtle blur. This is an early unfinished attempt at a x1 Lite model designed specifically for enhancing detail on skin diffuse textures of 3d characters. Even in its current state it works quite well.
+
|-
  
Best suited for uncompressed or cleaned textures - otherwise it may just enhance any existing compression artefacts too. The training set included faces, body parts, eyes and hair in a variety of skin types and tones so it should work well on most related diffuse textures.
+
| [https://drive.google.com/file/d/12v_2OwrB_rr6_0tMshHaHgD_XWRNgIyF/view 1x_BSChroma]
 
+
| |[[User:BlackScout|BlackScout]]
However it's not just limited to skin, many other images and textures can be enhanced with this model. The results are subtle, so run multiple times if desired.
+
| 1x
| 50/50 Interpolation of DIV2K-Lite and SpongeBC1-Lite
+
| [https://www.gnu.org/licenses/gpl-3.0.en.html GNU GPLv3]
 +
| ESRGAN
 +
| Sharpen
 +
| 2020-02-29
 +
| ChromaSharpen - makes the colors slightly more vibrant with a sideeffect of possibly adding Chromatic Abberation to the image. I am not sure about the usage of this model on real case scenarios but anything blurry or with fuzzy colors could work
 +
| none
 
|  
 
|  
 
|-
 
|-
  
 
+
| [https://drive.google.com/file/d/1JLyZZbnb06BddjNMXZ_CAyyoMpWCDHq0/view?usp=sharing/view 1x_3mChroma]
 +
| |[[User:3majsie1995|3majsie1995]]
 +
| 1x
 +
| [https://creativecommons.org/licenses/by-nc/4.0/ CC BY-NC 4.0]
 +
| ESRGAN
 +
| Chroma
 +
| 2023-01-04
 +
| Improves chromatic aberration. It's not as good as 1x_BSchroma, but I recommend you to test it!
 +
| 1xESRGAN
 +
| https://i.imgur.com/oSub859.png
 +
|-
 
|}
 
|}
  
 
+
===Cats===
===Foliage/Ground===
 
 
{| class="wikitable sortable"
 
{| class="wikitable sortable"
 
|-
 
|-
Line 3,385: Line 3,541:
 
! Purpose (Full)
 
! Purpose (Full)
 
! Pretrained_Model_G
 
! Pretrained_Model_G
! sample
+
! Sample
  
 
|-
 
|-
| [https://drive.google.com/drive/folders/1ZcMD3hy1YCm-hsrT6xYGvhboxSzhrJKg 1x_Plants_400000_G.pth]
+
| |[https://f002.backblazeb2.com/file/ESRGAN/_RELEASE/4x_cat_patch_325000_G.pth Cat_Patch]
| Muf
+
| |[[User:Twittman|Twittman]]
| 1x
+
| 4x
| Public Domain
+
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0] (?)
 
| ESRGAN
 
| ESRGAN
| bad upscale
+
| Cats
| 2021-06-23
+
| not on discord
| Images of plants, trees or other foliage upscaled with Photoshop Preserve Details 2.0. Sharpens and "subdivides" details and noise so it doesn't look upscaled.
+
|  
| 1x_NMKD-h264Texturize_500k.pth
+
| previous attempt
| https://imgsli.com/NTg2MDk https://imgsli.com/NTg2MDg https://imgsli.com/NTg2MTA
+
|  
 
|-
 
|-
 +
|}
  
| [https://drive.google.com/file/d/1dGmhHUPmb3lO9buX_Bt2nq97Nk5MCTb4/view Ground]
+
===Coins===
| tldr_coder
+
{| class="wikitable sortable"
 +
|-
 +
! Model Name
 +
! Author
 +
! Scale
 +
! License
 +
! Architecture
 +
! Purpose (short)
 +
! Date Posted
 +
! Purpose (Full)
 +
! Pretrained_Model_G
 +
! Sample
 +
 
 +
|-
 +
| [https://drive.google.com/file/d/1zC97RLUg5ejqb5-4Dj5q-SzdLPLiMozH/view 4x_Nickelfront]
 +
| BlackScout
 
| 4x
 
| 4x
| WTFPL
+
| GNU GPLv3
 
| ESRGAN
 
| ESRGAN
| Upscales ground textures
+
| Coins
| 2021-06-23
+
| 2020-02-13
| Dataset: Custom (Ground textures Google)
+
| Upscale coins. That's it. If you were mad at me because Nickelback doesn't make any sense. Now you have the perfect solution to your problems. If you want to upscale nickels or anything with similar texture made out of metal, now you can. It works pretty well for a joke.
|  
+
| 4xESRGAN
 
|  
 
|  
 
|-
 
|-
 
|}
 
|}
  
===Video Games===
+
===Faces===
====Game Screenshots====
 
 
{| class="wikitable sortable"
 
{| class="wikitable sortable"
 
|-
 
|-
Line 3,426: Line 3,597:
 
! Purpose (Full)
 
! Purpose (Full)
 
! Pretrained_Model_G
 
! Pretrained_Model_G
! sample
+
! Sample
  
 
|-
 
|-
| [https://drive.google.com/file/d/1DbbM4saRpJu0dmgvccQFwfF4zI2rjiQ_ Minepack]
+
| [https://de-next.owncube.com/index.php/s/PF5GTJ9jix4Log8 4x_Faces_N_250000.pth]
| |[[User:BlackScout|BlackScout]]
+
| |[[User:Twittman|Twittman]]
 
| 4x
 
| 4x
| CC BY-NC-SA
+
|  
 
| ESRGAN
 
| ESRGAN
| Upscaling pack.png
+
| Faces
| 2020-01-26
+
| 2019-07-18
| Upscales Minecraft screenshots by 4x. May suffer from haloing, weird patterns on blocks and JPEG-like artifacts.
+
| Upscales images of faces of different scales, sometimes resulting in monsters.
| 4xESRGAN.pth
+
| 4xESRGAN
 
|  
 
|  
 
|-
 
|-
  
 
|-
 
|-
| [https://drive.google.com/drive/u/0/folders/1zM07-eMja0PISoUz0dWGqx8b_YspU-gl 4x_Link]
+
| [https://de-next.owncube.com/index.php/s/PF5GTJ9jix4Log8 4x_Faces_04_N_180000_G]
| Fielran#1024
+
| |[[User:Twittman|Twittman]]
 
| 4x
 
| 4x
| GNU GPL3
+
|  
 
| ESRGAN
 
| ESRGAN
| Chainmail game textures. Alternatively, it can be used to turn plain images into chainmail.
+
| Faces
| 2022-10-06
+
| 2019-07-23
| A highly generative model for chainmail game textures. Works fairly well on items without visible chainmail texture (ie just a grey area) or items with clear chainmail texture, less well on items with present but poorly defined chainmail texture. I'm not 100% happy with it but it is still an improvement over existing models in enough situations to be worth releasing.
+
| Upscale faces both pixelized and real
| 4x_Nickelfront_14000G.pth
+
| 4x_Faces_N_250000.pth
| https://cdn.discordapp.com/attachments/579685650824036387/1027685415756509265/4x_Link_500000_G-TestImage-0x9F6EFFE0.png https://cdn.discordapp.com/attachments/579685650824036387/1027685416180121722/4x_Link_500000_G-TestImage-CMMM_Knight_Templar_ID2_DiffuseMap.png https://cdn.discordapp.com/attachments/579685650824036387/1027685416528261203/4x_Link_500000_G-TestImage-Crest_Venetian_DiffuseMap.png
+
|  
 
|-
 
|-
  
 
|-
 
|-
| [https://mega.nz/folder/zRYh3SII#QIm6T-rzhxjBLeYF1zSDpg 4x-FatePlus-lite]
+
| [https://de-next.owncube.com/index.php/s/mZRG4HB3KdP2iP6 Face-Ality V1 (4x_Fatality_Faces_310000_G.pth)]
| |[[User:Kim2091|Kim2091]]
+
| |[[User:Twittman|Twittman]]
 
| 4x
 
| 4x
| CC BY-NC-SA 4.0
+
| [https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0] (?)
| ESRGAN-lite
+
| ESRGAN
| Anime PSP games, Fate Extra
+
| Faces
| 2022-07-03
+
|  
| This model was trained as a favor to Demon and the Fate Extra community. It leaves a nice grain on the images and upscales lines and details accurately without looking odd. This model works on most anime-style PSP games. Enjoy! It works best on content with dithering and quantization. '''NOTICE:''' I have included both NCNN and ONNX models to make upscaling easier if you rely on either of these. For NCNN, there's two versions. One is FP16 and the other is FP32. FP16 works best on RTX GPUs. Choose FP32 if in doubt about compatibility, or if FP16 doesn't work for you. To use ONNX, download chaiNNer and upscale through there with the ONNX nodes.
+
| Dataset: Custom (Faces)
| 4x-AnimeSharp-lite
+
| 4x_Faces_04_N_180000_G.pth
 
|  
 
|  
 
|-
 
|-
  
 
|-
 
|-
| [https://drive.google.com/file/d/15KhApV05F8nEqTa5ZjLE821rhe1k8VOg/view?usp=sharing 4x_GameAI_2.0] [https://drive.google.com/file/d/1-s7iSK9ivJaoyvTXwDZvMjuyGG72bTQA/view?usp=sharing 4x_GameAI_1.0]
+
| [https://1drv.ms/f/s!Aip-EMByJHY29BRdcYSnTpgx0LWd?e=nX92z1 4x_SmolFace_200k]
| Tal
+
| [[User:DinJerr|DinJerr]]
 
| 4x
 
| 4x
| WTFPL
+
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 
| ESRGAN
 
| ESRGAN
| PS2 textures, cartoonish and realistic game textures.
+
| Art/People
| 2022-02-27
+
| 2022-08-19
| This model is intended to mainly handle PS2 compression and a mixture of Realistic and cartoonish textures, it's not meant to be used for very low resolution textures such as item icons. Dataset: I used textures from, "(A Hat in Time, Kingdom Hearts 3 and World of Final Fantasy) for GameAI_2.0" "(Skyrim, The Witcher 3, Resident Evil 4, Final Fantasy XV and Tales of Vesperia ) for GameAI_1.0" Dataset_size: 21.9k to 70k textures , up to 1024x1024 per texture
+
| A sharp upscaler trained specifically for small sprite faces. Does not blend, so avoid using on painted/photo portraits unless you were trying to retain more of the outlines somehow. The _clean version has denoising/dedithering training on top of it.
| 4x_GameAI_1.0
+
| 4x_NMKD-UltraYandere_300k
| https://cdn.discordapp.com/attachments/579685650824036387/947610558981632040/2ea384c8dc05593e-55207ec648ad52a4-00006213_2-comparison.png https://cdn.discordapp.com/attachments/579685650824036387/947610559497510922/65aac3d41ce0297c-5ff21b7095173a35-00006213_2-comparison.png https://cdn.discordapp.com/attachments/579685650824036387/947610560005033984/a6518186b4e88a86-ddbb8bd9666cd975-00006213_2-comparison.png https://cdn.discordapp.com/attachments/579685650824036387/947610560369950740/ae6ae340a472b13d-2ade3b7d0cf4a09f-00005dd3_2-comparison.png
+
| SmolFace: https://imgsli.com/MTIxNjU1/8/9
 +
SmolFace_clean: [https://1drv.ms/i/s!Aip-EMByJHY2-TNPYE5i1apNxRPn?e=B4tBsW Sample 1] [https://1drv.ms/i/s!Aip-EMByJHY2gYVZac1TWZP2A52_tw?e=Wrfnjo Sample 2] [https://1drv.ms/i/s!Aip-EMByJHY2-TSVRq30c_KUe1Mc?e=GE1TO9 Comparison with anime models]
 
|-
 
|-
  
| [https://drive.google.com/file/d/1Xe1JorGFJzPClR9tUuGmpuMyLhTA-d2T/view?usp=sharing 4xPackCraft_v4]
+
|-
| Joey
+
| [https://1drv.ms/u/s!Aip-EMByJHY27CHg3-1Uue5KgbtK?e=6k0COU BigFace_v3, BigFace_V3_Blend, BigFace_V3_Clear]
 +
| [[User:DinJerr|DinJerr]]
 
| 4x
 
| 4x
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
+
|[https://creativecommons.org/licenses/by-nc/4.0/ CC BY-NC 4.0]
 
| ESRGAN
 
| ESRGAN
| Upscaling pack.png
+
| Art/People
| 2020-09-05
+
| 2020-07-22
| Designed to upscale one specific minecraft screenshot. Results on dissimilar screenshots may be poor.
+
| Pixel art upscaler for faces drawn in digital painting style. Best for game portraits with multiple shades (not cel-shaded).
| 4xESRGAN/previous versions
+
| 4x_BigFace
| https://cdn.discordapp.com/attachments/436411385673547787/750481717307113543/pack.png
+
|
 
|-
 
|-
  
 
|-
 
|-
| [https://drive.google.com/drive/folders/1XsOFqGx_cGMWQTBuE9K9smwULg_wiaWc?usp=sharing MinecraftAlphaUpscaler with Good data]
+
| [https://drive.google.com/open?id=1OyOJIW224hBhb-aTCbuUQb0qzKmE4oH6 TGHQFace8x]
| Washed Up
+
| tg
 +
| 8x
 +
| GNU GPL3
 +
| ESRGAN
 +
| Faces
 +
| 2019-11-27
 +
| Upscales blurry 128px faces, usefull for enhancing that someone in a small picture.
 +
| 8xESRGAN
 +
| https://imgsli.com/OTM1NQ
 +
|-
 +
 
 +
| [https://1drv.ms/u/s!Aip-EMByJHY211Vcr3fF39ZOaPk-?e=BQajTd FArtFace]
 +
| |[[User:DinJerr|DinJerr]]
 
| 4x
 
| 4x
| [https://www.gnu.org/licenses/gpl-3.0.en.html GNU GPLv3]
+
|[https://creativecommons.org/licenses/by-nc/4.0/ CC BY-NC 4.0]
 
| ESRGAN
 
| ESRGAN
| Upscaling pack.png
+
| Faces
| 2020-01-29
+
| 2019-11-28
| Designed to upscale one specific minecraft screenshot. Results on dissimilar screenshots may be poor.
+
| Rather old pixel art face upscaler. I suggest using BigFace_v3 or SmolFace instead. Sometimes produces better results though.
| RRDB_PNSR_x4.pth
+
|  
 
|  
 
|  
 
|-
 
|-
  
|-
+
| [https://drive.google.com/file/d/14daAyYuoX-nZ_Lo1U6oekbFqrZTQlQkG/view?usp=drivesdk 4x_BS_SbeveHarvey]
| [https://drive.google.com/file/d/1MkqAuELmbhrEGcD55IUJhDFFNniOmNWN ScreenBooster V2]
 
 
| |[[User:BlackScout|BlackScout]]
 
| |[[User:BlackScout|BlackScout]]
 
| 4x
 
| 4x
| CC BY-NC-SA
+
| [https://www.gnu.org/licenses/gpl-3.0.en.html GNU GPLv3]
 
| ESRGAN
 
| ESRGAN
| CGI
+
| Faces
| 2020-01-29
+
| 2020-07-24
| Game Screenshots
+
| Upscale Steve Harvey, but maybe other things, somehow??
| none
+
| 4xESRGAN
 
|  
 
|  
 
|-
 
|-
  
 
|-
 
|-
| [https://drive.google.com/file/d/1JI8VZSSHpH9u7HzE8MvuNG351Dgn4jwU/view?usp=sharing BS_ScreenBooster_SPSR]
+
| [https://drive.google.com/file/d/19ICMKNuS4PbhmtA7Be9lUE6v8NHDDEzP/view Face Focus]
| |[[User:BlackScout|BlackScout]]
+
| |[[User:LyonHrt|LyonHrt]]
 
| 4x
 
| 4x
| CC BY-NC-SA
+
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
| SPSR
+
| ESRGAN
| CGI
+
| Faces
| 2020-07-18
+
| not on discord
| This model is designed to upscale game screenshots (3D Games) by 4 times. The SPSR version is an improvement over the ESRGAN based V2.
+
| Face De-blur - slightly out of focus / blurred images of faces. It is aimed at faces / hair
| Screenbooster V2
+
| 4xPSNR
 +
|  
 +
|-
 +
 
 +
|-
 +
| [https://drive.google.com/file/d/1OyOJIW224hBhb-aTCbuUQb0qzKmE4oH6/view TGHQFace8x]
 +
| |[[User:Torrentguy|Torrentguy]]
 +
| 8x
 +
| [https://www.gnu.org/licenses/gpl-3.0.en.html [https://www.gnu.org/licenses/gpl-3.0.en.html GNU GPLv3]]
 +
| ESRGAN
 +
| Face Upscaling
 +
|
 +
|
 +
| 8xESRGAN
 
|  
 
|  
 
|-
 
|-
 
|}
 
|}
  
====Normal Map/Bump Map Generation====
+
===Skin===
You may want to use this instead: https://github.com/JoeyBallentine/Material-Map-Generator
 
 
 
 
{| class="wikitable sortable"
 
{| class="wikitable sortable"
 
|-
 
|-
Line 3,546: Line 3,741:
 
! Purpose (Full)
 
! Purpose (Full)
 
! Pretrained_Model_G
 
! Pretrained_Model_G
! sample
+
! Sample
 +
 
  
 
|-
 
|-
| [https://u.pcloud.link/publink/show?code=kZEdNHXZ3bpVwUoge94d9igFWxbjQQrcbpny 1x_NormalMapGenerator-CX-Lite]
+
| [https://drive.google.com/drive/folders/1VkT6tpbCPn2gKZYPtawDJGMpLg6EyRpO?usp=sharing x1_ITF_SkinDiffDetail_Lite_v1]
| Joey
+
| intheflesh#3116
 
| 1x
 
| 1x
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
+
| CC BY-NC-SA 4.0
| ESRGAN Lite [nf=32 nb=12]
+
| ESRGAN
| Map Generation - normal maps
+
| Skin Upscaling
| 2020-11-04
+
|  
| Generating normal maps from textures
+
| Adds plausible high frequency detail and removes subtle blur. This is an early unfinished attempt at a x1 Lite model designed specifically for enhancing detail on skin diffuse textures of 3d characters. Even in its current state it works quite well.
| 1x_DIV2K-Lite_450k.pth
+
 
|
+
Best suited for uncompressed or cleaned textures - otherwise it may just enhance any existing compression artefacts too. The training set included faces, body parts, eyes and hair in a variety of skin types and tones so it should work well on most related diffuse textures.
|-
 
  
| [https://u.pcloud.link/publink/show?code=kZecVzXZEO6fAyO6w25YDc43BkSVHRgFdOwX 1x_FrankenMapGenerator-CX-Lite]
+
However it's not just limited to skin, many other images and textures can be enhanced with this model. The results are subtle, so run multiple times if desired.
| Joey
+
| 50/50 Interpolation of DIV2K-Lite and SpongeBC1-Lite
| 1x
 
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 
| ESRGAN Lite [nf=32 nb=12]
 
| Map Generation - roughness and displacement maps
 
| 2020-11-07
 
| This model generates "Franken Maps" (named after Frankenstein), which is a custom material map combination I made. Basically, the Red channel of RGB is just the texture converted to grayscale, the Green channel is the roughness map, and the Blue channel is the displacement map. I had to do this to get around the current limitation of CX loss where it requires a 3 channel output (otherwise I would have just made a 2 channel model, or separate single channel models). As of right now the channels need to be manually split from each other but I will be making a tool for doing this automatically in the coming days.
 
| 1x_DIV2K-Lite_450k.pth
 
 
|  
 
|  
 
|-
 
|-
  
| [https://drive.google.com/file/d/1uksv1uyyZjQcU4567OrfU0w2NECcG8dY/view 1x_normals_generator_general_215k]
 
| [[User:LyonHrt|LyonHrt]]
 
| 1x
 
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 
| ESRGAN
 
| Map Generation - Normal Maps
 
|
 
| This model generates "Franken Maps" (named after Frankenstein), which is a custom material map combination I made. Basically, the Red channel of RGB is just the texture converted to grayscale, the Green channel is the roughness map, and the Blue channel is the displacement map. I had to do this to get around the current limitation of CX loss where it requires a 3 channel output (otherwise I would have just made a 2 channel model, or separate single channel models). As of right now the channels need to be manually split from each other but I will be making a tool for doing this automatically in the coming days.
 
| none (no interpolation)
 
|
 
|-
 
  
 
|}
 
|}
  
====Video Game Textures====
+
 
 +
===Foliage/Ground===
 
{| class="wikitable sortable"
 
{| class="wikitable sortable"
 
|-
 
|-
Line 3,599: Line 3,777:
 
! Purpose (Full)
 
! Purpose (Full)
 
! Pretrained_Model_G
 
! Pretrained_Model_G
! sample
+
! Sample
 
 
  
 
|-
 
|-
| [https://1drv.ms/u/s!AiWox1lAWLoTg1bUeJouJNOFZ_Jj Fallout Weapons (Fallout 4 Weapons?)]
+
| [https://drive.google.com/drive/folders/1ZcMD3hy1YCm-hsrT6xYGvhboxSzhrJKg 1x_Plants_400000_G.pth]
| Bob
+
| Muf
| 4x
+
| 1x
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
+
| Public Domain
 
| ESRGAN
 
| ESRGAN
| Game textures
+
| bad upscale
| 2019-08-05
+
| 2021-06-23
| Video game textures, mostly metal rusty, clean or painted
+
| Images of plants, trees or other foliage upscaled with Photoshop Preserve Details 2.0. Sharpens and "subdivides" details and noise so it doesn't look upscaled.
| Manga109Attempt
+
| 1x_NMKD-h264Texturize_500k.pth
|  
+
| https://imgsli.com/NTg2MDk https://imgsli.com/NTg2MDg https://imgsli.com/NTg2MTA
 
|-
 
|-
  
| [https://1drv.ms/u/s!AiWox1lAWLoThANuXjxIR-hqA6os Fallout Weapons V2]
+
| [https://drive.google.com/file/d/1dGmhHUPmb3lO9buX_Bt2nq97Nk5MCTb4/view Ground]
| Bob
+
| tldr_coder
 
| 4x
 
| 4x
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
+
| WTFPL
 
| ESRGAN
 
| ESRGAN
| Game textures
+
| Upscales ground textures
| 2019-08-23
+
| 2021-06-23
| Video game textures, mostly metal rusty, clean or painted
+
| Dataset: Custom (Ground textures Google)
| Fallout 4 Weapons
+
|  
 
|  
 
|  
 
|-
 
|-
 +
|}
  
 +
===Video Games===
 +
====Game Screenshots====
 +
{| class="wikitable sortable"
 +
|-
 +
! Model Name
 +
! Author
 +
! Scale
 +
! License
 +
! Architecture
 +
! Purpose (short)
 +
! Date Posted
 +
! Purpose (Full)
 +
! Pretrained_Model_G
 +
! Sample
  
| [https://drive.google.com/file/d/1y9nYjuy8vG7_LHpveYTj0ihK_6fE2XKF/view?usp=sharing 4x_ThiefGoldMod_100000]
+
|-
| Akven
+
| [https://drive.google.com/file/d/1DbbM4saRpJu0dmgvccQFwfF4zI2rjiQ_ Minepack]
 +
| |[[User:BlackScout|BlackScout]]
 
| 4x
 
| 4x
| no idea
+
| CC BY-NC-SA
 
| ESRGAN
 
| ESRGAN
| Game textures
+
| Upscaling pack.png
| 2020-05-15
+
| 2020-01-26
| Version of the previous model but based on Manga109 pretrained model and with slightly different dataset. Sometimes gives better results especially for wood and metal, sometimes worse. Sometime generates the same dotted artifacts on very bright/white images.
+
| Upscales Minecraft screenshots by 4x. May suffer from haloing, weird patterns on blocks and JPEG-like artifacts.
| 4x_Manga109Attempt
+
| 4xESRGAN.pth
| https://imgsli.com/MTYzNjI
+
|  
 
|-
 
|-
  
 
+
|-
| [https://drive.google.com/file/d/1L8reADxi5Nt-8Tki0oY5jkQlv6kJg6w4/view?usp=sharing 4x_ThiefGold_110000]
+
| [https://drive.google.com/drive/u/0/folders/1zM07-eMja0PISoUz0dWGqx8b_YspU-gl 4x_Link]
| Akven
+
| Fielran#1024
 
| 4x
 
| 4x
| no idea
+
| GNU GPL3
 
| ESRGAN
 
| ESRGAN
| Game textures
+
| Chainmail game textures. Alternatively, it can be used to turn plain images into chainmail.
| 2020-05-15
+
| 2022-10-06
| Various game textures. Primary wood, metal, stone
+
| A highly generative model for chainmail game textures. Works fairly well on items without visible chainmail texture (ie just a grey area) or items with clear chainmail texture, less well on items with present but poorly defined chainmail texture. I'm not 100% happy with it but it is still an improvement over existing models in enough situations to be worth releasing.
| RRDB_ESRGAN_x4
+
| 4x_Nickelfront_14000G.pth
| https://imgsli.com/MTYzNjI
+
| [https://cdn.discordapp.com/attachments/579685650824036387/1027685415756509265/4x_Link_500000_G-TestImage-0x9F6EFFE0.png Sample 1]
 +
 
 +
[https://cdn.discordapp.com/attachments/579685650824036387/1027685416180121722/4x_Link_500000_G-TestImage-CMMM_Knight_Templar_ID2_DiffuseMap.png Sample 2]
 +
 
 +
[https://cdn.discordapp.com/attachments/579685650824036387/1027685416528261203/4x_Link_500000_G-TestImage-Crest_Venetian_DiffuseMap.png Sample 3]
 
|-
 
|-
  
| [https://mega.nz/folder/HNh0hLzD#vUuFhp2ZAeNi_5shZfCC0g 4x-VolArt]
+
|-
 +
| [https://mega.nz/folder/zRYh3SII#QIm6T-rzhxjBLeYF1zSDpg 4x-FatePlus-lite]
 
| |[[User:Kim2091|Kim2091]]
 
| |[[User:Kim2091|Kim2091]]
 
| 4x
 
| 4x
 
| CC BY-NC-SA 4.0
 
| CC BY-NC-SA 4.0
| ESRGAN
+
| ESRGAN-lite
| Game textures/Art
+
| Anime PSP games, Fate Extra
| 2021-09-18
+
| 2022-07-03
| This model upscales artwork for the game Volfoss (2001). The model peaked in quality very quickly. The NR model removes most noise, but has the downside of removing transparent portions. Use the main model in most cases.
+
| This model was trained as a favor to Demon and the Fate Extra community. It leaves a nice grain on the images and upscales lines and details accurately without looking odd. This model works on most anime-style PSP games. Enjoy! It works best on content with dithering and quantization. '''NOTICE:''' I have included both NCNN and ONNX models to make upscaling easier if you rely on either of these. For NCNN, there's two versions. One is FP16 and the other is FP32. FP16 works best on RTX GPUs. Choose FP32 if in doubt about compatibility, or if FP16 doesn't work for you. To use ONNX, download chaiNNer and upscale through there with the ONNX nodes.
| 4xESRGAN
+
| 4x-AnimeSharp-lite
| [https://cdn.discordapp.com/attachments/888869052867543060/888869070928248852/unknown.png Sample 1] [https://cdn.discordapp.com/attachments/888869052867543060/888869784169619496/unknown.png Sample 2]
+
|
 +
|-
 +
 
 
|-
 
|-
 +
| [https://drive.google.com/file/d/15KhApV05F8nEqTa5ZjLE821rhe1k8VOg/view?usp=sharing 4x_GameAI_2.0] [https://drive.google.com/file/d/1-s7iSK9ivJaoyvTXwDZvMjuyGG72bTQA/view?usp=sharing 4x_GameAI_1.0]
 +
| Tal
 +
| 4x
 +
| WTFPL
 +
| ESRGAN
 +
| PS2 textures, cartoonish and realistic game textures.
 +
| 2022-02-27
 +
| This model is intended to mainly handle PS2 compression and a mixture of Realistic and cartoonish textures, it's not meant to be used for very low resolution textures such as item icons. Dataset: I used textures from, "(A Hat in Time, Kingdom Hearts 3 and World of Final Fantasy) for GameAI_2.0" "(Skyrim, The Witcher 3, Resident Evil 4, Final Fantasy XV and Tales of Vesperia ) for GameAI_1.0" Dataset_size: 21.9k to 70k textures , up to 1024x1024 per texture
 +
| 4x_GameAI_1.0
 +
| [https://cdn.discordapp.com/attachments/579685650824036387/947610558981632040/2ea384c8dc05593e-55207ec648ad52a4-00006213_2-comparison.png Sample 1]
  
 +
[https://cdn.discordapp.com/attachments/579685650824036387/947610559497510922/65aac3d41ce0297c-5ff21b7095173a35-00006213_2-comparison.png Sample 2]
  
| [https://drive.google.com/file/d/1Ful_AALrK-gxSwsTeMkzvwX5hh1ah8X8/view?usp=sharing 2xFaithfulSPSR]
+
[https://cdn.discordapp.com/attachments/579685650824036387/947610560005033984/a6518186b4e88a86-ddbb8bd9666cd975-00006213_2-comparison.png Sample 3]  
| Joey
+
 
| 2x
+
[https://cdn.discordapp.com/attachments/579685650824036387/947610560369950740/ae6ae340a472b13d-2ade3b7d0cf4a09f-00005dd3_2-comparison.png Sample 4]
| CC BY-NC-SA
 
| SPSR
 
| Pixel art
 
| 2020-07-19
 
| Mainly just a test for SPSR. Seems to work better than the original 2xFaithful32_1316 that I used as a pretrained, even though it uses the same dataset.
 
| 2xFaithful32_1316
 
|
 
 
|-
 
|-
  
| [https://u.pcloud.link/publink/show?code=kZUMKFXZUG6BpzHR6wFGack19pvWgbQQ9OA7 2x_FakeFaith-Lite]
+
| [https://drive.google.com/file/d/1Xe1JorGFJzPClR9tUuGmpuMyLhTA-d2T/view?usp=sharing 4xPackCraft_v4]
 
| Joey
 
| Joey
| 2x
+
| 4x
 
|[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]
| ESRGAN (lite)
+
| ESRGAN
| Pixel Art
+
| Upscaling pack.png
| 2020-10-12
+
| 2020-09-05
| An attempt at recreating the "faithful" style without using the faithful dataset -- aka keeping the "pixel art" style of pixel art.
+
| Designed to upscale one specific minecraft screenshot. Results on dissimilar screenshots may be poor.
| 2x_Faithful-Lite
+
| 4xESRGAN/previous versions
|  
+
| [https://cdn.discordapp.com/attachments/436411385673547787/750481717307113543/pack.png Sample]
 
|-
 
|-
  
| [https://u.pcloud.link/publink/show?code=XZHawFXZhqD7dF4qtYHnqJVPnQd0ey0vW1JX 2x_Faithful-Lite]
+
|-
| Joey
+
| [https://drive.google.com/drive/folders/1XsOFqGx_cGMWQTBuE9K9smwULg_wiaWc?usp=sharing MinecraftAlphaUpscaler with Good data]
| 2x
+
| Washed Up
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
+
| 4x
| ESRGAN (lite)
+
| [https://www.gnu.org/licenses/gpl-3.0.en.html GNU GPLv3]
| Pixel Art
+
| ESRGAN
| 2020-10-12
+
| Upscaling pack.png
| A "lite" model version of my Faithful model
+
| 2020-01-29
| none
+
| Designed to upscale one specific minecraft screenshot. Results on dissimilar screenshots may be poor.
 +
| RRDB_PNSR_x4.pth
 
|  
 
|  
 
|-
 
|-
  
 
|-
 
|-
| [https://github.com/Venomalia/HDcube/tree/main/v3#hdcube-3 4x_HDCube3]
+
| [https://drive.google.com/file/d/1MkqAuELmbhrEGcD55IUJhDFFNniOmNWN ScreenBooster V2]
| Venomalia
+
| |[[User:BlackScout|BlackScout]]
 
| 4x
 
| 4x
| CC BY-NC-SA 4.0
+
| CC BY-NC-SA
 
| ESRGAN
 
| ESRGAN
| Gamecube and Wii textures.
+
| CGI
| 2022-12-13
+
| 2020-01-29
| It can be used for all image formats supported by Gamecube and Wii hardware and can remove its typical artifacts like CMPR Block Compression (DXT1 algorithm, also known as BC1), color palette errors, color reduction up to 8bit color depth and 1bit alpha depth.
+
| Game Screenshots
| 4x_HDcube2
+
| none
| https://imgsli.com/MTM5ODE0/0/1 https://github.com/Venomalia/HDcube/tree/main/v3#image-comparison
+
|  
 
|-
 
|-
  
 
|-
 
|-
| [https://github.com/Venomalia/HDcube/tree/main/v1#hdcube 4x_HDCube]
+
| [https://drive.google.com/file/d/1JI8VZSSHpH9u7HzE8MvuNG351Dgn4jwU/view?usp=sharing BS_ScreenBooster_SPSR]
| Venomalia
+
| |[[User:BlackScout|BlackScout]]
 
| 4x
 
| 4x
| CC0-1.0
+
| CC BY-NC-SA
| ESRGAN
+
| SPSR
| Gamecube and Wii textures.
+
| CGI
| 2022-05-10
+
| 2020-07-18
| Gamecube and Wii textures (mainly DXT and 8bit color compression). Is good for preserving fine details without affecting the original style too much, it is not suitable for pixel art, small icons and text under 16 pixel.
+
| This model is designed to upscale game screenshots (3D Games) by 4 times. The SPSR version is an improvement over the ESRGAN based V2.
| 4x_NMKD Siax
+
| Screenbooster V2
| https://imgsli.com/MTA3NDQ3 https://slow.pics/c/rmQsQ6yb
+
|  
 
|-
 
|-
 +
|}
  
| [https://u.pcloud.link/publink/show?code=XZTAmsXZzEO8tMJaYAF1MrKKnmOCKu7Ghedy 1x_DXTless_SourceEngine_170000_G]
+
====Normal Map/Bump Map Generation====
| Xeller
+
You may want to use this instead: https://github.com/JoeyBallentine/Material-Map-Generator
| 1x
+
 
| WTFPL
+
{| class="wikitable sortable"
| ESRGAN
 
| DTX5 Compression
 
| 2021-04-10
 
| This model is made for Source Engine textures. It tries to remove compression artifacts such as blockyness, discoloration, green tint. It does pretty well on a lot of things, realistic stuff as well, but it was mostly made to work on TF2 textures. It's made to keep as much detail as possible, without any unnecessary denoising/sharpening. Huge thanks to Twittman for assisting me along this journey.
 
| 1x_Saiyajin_DeJpeg_300000_G
 
| https://imgsli.com/NDk1NTk https://imgsli.com/NDk1NjA https://imgsli.com/NDk1NjE https://imgsli.com/NDk1NjI
 
 
|-
 
|-
 +
! Model Name
 +
! Author
 +
! Scale
 +
! License
 +
! Architecture
 +
! Purpose (short)
 +
! Date Posted
 +
! Purpose (Full)
 +
! Pretrained_Model_G
 +
! Sample
  
| [https://drive.google.com/file/d/1HIBRKFs7s-XhpN1p7rwAWcMTHwFRqn3n/view Trixie]
+
|-
| |[[User:LyonHrt|LyonHrt]]
+
| [https://u.pcloud.link/publink/show?code=kZEdNHXZ3bpVwUoge94d9igFWxbjQQrcbpny 1x_NormalMapGenerator-CX-Lite]
| 4x
+
| Joey
 +
| 1x
 
|[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]
| ESRGAN
+
| ESRGAN Lite [nf=32 nb=12]
| Faces / Game Textures
+
| Map Generation - normal maps
| not on discord
+
| 2020-11-04
| 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
+
| Generating normal maps from textures
| none
+
| 1x_DIV2K-Lite_450k.pth
 +
|
 +
|-
 +
 
 +
| [https://u.pcloud.link/publink/show?code=kZecVzXZEO6fAyO6w25YDc43BkSVHRgFdOwX 1x_FrankenMapGenerator-CX-Lite]
 +
| Joey
 +
| 1x
 +
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 +
| ESRGAN Lite [nf=32 nb=12]
 +
| Map Generation - roughness and displacement maps
 +
| 2020-11-07
 +
| This model generates "Franken Maps" (named after Frankenstein), which is a custom material map combination I made. Basically, the Red channel of RGB is just the texture converted to grayscale, the Green channel is the roughness map, and the Blue channel is the displacement map. I had to do this to get around the current limitation of CX loss where it requires a 3 channel output (otherwise I would have just made a 2 channel model, or separate single channel models). As of right now the channels need to be manually split from each other but I will be making a tool for doing this automatically in the coming days.
 +
| 1x_DIV2K-Lite_450k.pth
 
|  
 
|  
 
|-
 
|-
  
| [https://e.pcloud.link/publink/show?code=XZGmsZo1yNrot19O4R5SrPVD1GWpRy1MKX Skyrim Armory]
+
| [https://drive.google.com/file/d/1uksv1uyyZjQcU4567OrfU0w2NECcG8dY/view 1x_normals_generator_general_215k]
| |[[User:Alsa|Alsa]]
+
| [[User:LyonHrt|LyonHrt]]
| 4x
+
| 1x
 
|[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]
 
| ESRGAN
 
| ESRGAN
| Game textures/equipment
+
| Map Generation - Normal Maps
| not on discord
 
 
|  
 
|  
| Manga109Attempt
+
| This model generates "Franken Maps" (named after Frankenstein), which is a custom material map combination I made. Basically, the Red channel of RGB is just the texture converted to grayscale, the Green channel is the roughness map, and the Blue channel is the displacement map. I had to do this to get around the current limitation of CX loss where it requires a 3 channel output (otherwise I would have just made a 2 channel model, or separate single channel models). As of right now the channels need to be manually split from each other but I will be making a tool for doing this automatically in the coming days.
 +
| none (no interpolation)
 
|  
 
|  
 
|-
 
|-
  
 +
|}
  
| [https://drive.google.com/file/d/1RjVWNUFtVyz4ykqu6Pm9wmO03pPEBGqg/view Skyrim Misc]
+
====Video Game Textures====
| Deorder
+
{| class="wikitable sortable"
| 4x
 
| CC0
 
| ESRGAN
 
| Game textures
 
| not on discord
 
| Skyrim Diffuse Textures
 
| ?
 
|
 
|-
 
 
 
| [https://drive.google.com/file/d/1RjVWNUFtVyz4ykqu6Pm9wmO03pPEBGqg/view Skyrim Wood]
 
| Laeris
 
| 4x
 
|
 
| ESRGAN
 
| Game textures
 
| not on discord
 
| Wood
 
| ?
 
|
 
 
|-
 
|-
 +
! Model Name
 +
! Author
 +
! Scale
 +
! License
 +
! Architecture
 +
! Purpose (short)
 +
! Date Posted
 +
! Purpose (Full)
 +
! Pretrained_Model_G
 +
! Sample
 +
  
 
|-
 
|-
| [https://mega.nz/folder/fcp2SLKC#tF3vG7zjvfpnXK7OLUaFRA 4x-SkyrimTexV2.1 and V2_Fabric]
+
| [https://1drv.ms/u/s!AiWox1lAWLoTg1bUeJouJNOFZ_Jj Fallout Weapons (Fallout 4 Weapons?)]
| [[User:Kim2091|Kim2091]]
+
| Bob
 
| 4x
 
| 4x
| CC BY-NC-SA 4.0
+
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 
| ESRGAN
 
| ESRGAN
 
| Game textures
 
| Game textures
| 2021-09-25
+
| 2019-08-05
| This model set upscales most Skyrim textures. I hope it helps 🙂 The base model (2.1) works well on stone, wood, metals, and most other textures. The Fabric model is a 50/50 interpolation with a stronger iteration of this model and my Fabric model. As you can tell by the name, it's intended for Fabric textures. If you want to upscale Alpha textures, use a model dedicated to it. There's an INFO file in the folder, it just explains the extra models
+
| Video game textures, mostly metal rusty, clean or painted
| 4xESRGAN
+
| Manga109Attempt
| [https://cdn.discordapp.com/attachments/891300021453070387/891392169926098974/unknown.png Sample 1] [https://cdn.discordapp.com/attachments/891393393849147423/891397870266224731/unknown.png Sample 2] [https://cdn.discordapp.com/attachments/891300021453070387/891395469387890738/unknown.png Sample 3] [https://cdn.discordapp.com/attachments/891393393849147423/891398051325960202/unknown.png Sample 4]
+
|  
 
|-
 
|-
  
| [https://drive.google.com/file/d/12fR-pWw6YL2ZbS6EDvrAeOVoQHWkWl5L/view Forest]
+
| [https://1drv.ms/u/s!AiWox1lAWLoThANuXjxIR-hqA6os Fallout Weapons V2]
| |[[User:LyonHrt|LyonHrt]]
+
| Bob
 
| 4x
 
| 4x
 
|[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]
 
| ESRGAN
 
| ESRGAN
 
| Game textures
 
| Game textures
| not on discord
+
| 2019-08-23
| Wood / Leaves
+
| Video game textures, mostly metal rusty, clean or painted
| none
+
| Fallout 4 Weapons
 
|  
 
|  
 
|-
 
|-
  
| [https://mega.nz/folder/tFhl0L5b#CK4WLHYxjX7KFXU1PLNDuQ Morrowind Mixed]
+
 
| mkultra
+
| [https://drive.google.com/file/d/1y9nYjuy8vG7_LHpveYTj0ihK_6fE2XKF/view?usp=sharing 4x_ThiefGoldMod_100000]
 +
| Akven
 
| 4x
 
| 4x
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
+
| no idea
 
| ESRGAN
 
| ESRGAN
 
| Game textures
 
| Game textures
| not on discord
+
| 2020-05-15
| Morrowind Mod Textures
+
| Version of the previous model but based on Manga109 pretrained model and with slightly different dataset. Sometimes gives better results especially for wood and metal, sometimes worse. Sometime generates the same dotted artifacts on very bright/white images.
| 4xESRGAN
+
| 4x_Manga109Attempt
|  
+
| https://imgsli.com/MTYzNjI
 
|-
 
|-
  
| [https://mega.nz/folder/dVQxxITT#Bsloq9G_Xh5tsDMKDQuTTA Morrowind 2.0]
+
 
| mkultra
+
| [https://drive.google.com/file/d/1L8reADxi5Nt-8Tki0oY5jkQlv6kJg6w4/view?usp=sharing 4x_ThiefGold_110000]
 +
| Akven
 
| 4x
 
| 4x
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
+
| no idea
 
| ESRGAN
 
| ESRGAN
 
| Game textures
 
| Game textures
| not on discord
+
| 2020-05-15
| Morrowind Mod Textures
+
| Various game textures. Primary wood, metal, stone
| 4xESRGAN
+
| RRDB_ESRGAN_x4
|  
+
| https://imgsli.com/MTYzNjI
 
|-
 
|-
  
| [https://drive.google.com/file/d/1xPzqzTG5L6y5I0U3W0oQjexwGF5zfG32/view Map]
+
| [https://mega.nz/folder/HNh0hLzD#vUuFhp2ZAeNi_5shZfCC0g 4x-VolArt]
| |[[User:LyonHrt|LyonHrt]]
+
| |[[User:Kim2091|Kim2091]]
 
| 4x
 
| 4x
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
+
| CC BY-NC-SA 4.0
 
| ESRGAN
 
| ESRGAN
| Game textures
+
| Game textures/Art
| not on discord
+
| 2021-09-18
| Map / Old Paper with text
+
| This model upscales artwork for the game Volfoss (2001). The model peaked in quality very quickly. The NR model removes most noise, but has the downside of removing transparent portions. Use the main model in most cases.
| none
+
| 4xESRGAN
|  
+
| [https://cdn.discordapp.com/attachments/888869052867543060/888869070928248852/unknown.png Sample 1] [https://cdn.discordapp.com/attachments/888869052867543060/888869784169619496/unknown.png Sample 2]
 
|-
 
|-
  
  
|}
+
| [https://drive.google.com/file/d/1Ful_AALrK-gxSwsTeMkzvwX5hh1ah8X8/view?usp=sharing 2xFaithfulSPSR]
 
+
| Joey
==Video Restoration==
+
| 2x
===Video Compression===
+
| CC BY-NC-SA
 
+
| SPSR
{| class="wikitable sortable"
+
| Pixel art
 +
| 2020-07-19
 +
| Mainly just a test for SPSR. Seems to work better than the original 2xFaithful32_1316 that I used as a pretrained, even though it uses the same dataset.
 +
| 2xFaithful32_1316
 +
|  
 
|-
 
|-
! Model Name
 
! Author
 
! Scale
 
! License
 
! Architecture
 
! Purpose (short)
 
! Date Posted
 
! Purpose (Full)
 
! Pretrained_Model_G
 
! sample
 
  
|-
+
| [https://u.pcloud.link/publink/show?code=kZUMKFXZUG6BpzHR6wFGack19pvWgbQQ9OA7 2x_FakeFaith-Lite]
| [https://de-next.owncube.com/index.php/s/wKjLmYsq7M5JAmx 1x_cinepak_200000]
+
| Joey
| |[[User:Twittman|Twittman]]
+
| 2x
| 1x
+
|[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] (?)
+
| ESRGAN (lite)
| ESRGAN
+
| Pixel Art
| Compression
+
| 2020-10-12
| 2019-07-03
+
| An attempt at recreating the "faithful" style without using the faithful dataset -- aka keeping the "pixel art" style of pixel art.
| Removal of Compression such as Cinepak, msvideo1 and Roq
+
| 2x_Faithful-Lite
| none
 
 
|  
 
|  
 
|-
 
|-
  
|-
+
| [https://u.pcloud.link/publink/show?code=XZHawFXZhqD7dF4qtYHnqJVPnQd0ey0vW1JX 2x_Faithful-Lite]
| [https://de-next.owncube.com/index.php/s/wKjLmYsq7M5JAmx 1x_cinepak_alt.pth]
+
| Joey
| |buildist
+
| 2x
| 1x
+
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
|[https://www.gnu.org/licenses/gpl-3.0.en.html GNU GPLv3]
+
| ESRGAN (lite)
| ESRGAN
+
| Pixel Art
| Compression
+
| 2020-10-12
|  
+
| A "lite" model version of my Faithful model
| Cinepak (only)
 
 
| none
 
| none
 
|  
 
|  
Line 3,896: Line 4,105:
  
 
|-
 
|-
| [https://mega.nz/folder/bYxyWaZI#YQXKPVCSPF4jzHDZUcI27A DeBink v4/v5/v6]
+
| [https://github.com/Venomalia/HDcube/tree/main/v3#hdcube-3 4x_HDCube3]
| |[[User:Kim2091|Kim2091]]
+
| Venomalia
| 1x
+
| 4x
 
| CC BY-NC-SA 4.0
 
| CC BY-NC-SA 4.0
 
| ESRGAN
 
| ESRGAN
| Compression
+
| Gamecube and Wii textures.
| 8.6.21
+
| 2022-12-13
| This model removes early 2000s Bink and other compression artifacts. Works well on almost any image or video compression type.
+
| It can be used for all image formats supported by Gamecube and Wii hardware and can remove its typical artifacts like CMPR Block Compression (DXT1 algorithm, also known as BC1), color palette errors, color reduction up to 8bit color depth and 1bit alpha depth.
| none
+
| 4x_HDcube2
|  
+
| https://imgsli.com/MTM5ODE0/0/1
|-
 
  
|-
+
[https://github.com/Venomalia/HDcube/tree/main/v3#image-comparison Sample Gallery]
| [https://mega.nz/folder/6cxHzIxB#IN6ZgG-S0j3LjM0x5DEqew DeBink_Lite]
 
| |[[User:Kim2091|Kim2091]]
 
| 1x
 
| CC BY-NC-SA 4.0
 
| ESRGAN-lite
 
| Compression
 
| 2021-10-30
 
| This model was trained on lossless video frames of Metal Arms: Glitch in the System, compressed with Bink V1 for the LR frames. It's a lot more efficient than my first DeBink model, and also has less artifacts. It's not quite as robust, but the compression is barely noticeable in videos after processing with this.
 
| none
 
| [https://cdn.discordapp.com/attachments/903415274521374750/904129653374078976/XB_Demo_Mov-85600_G.mp4 Video Sample]
 
 
|-
 
|-
 
  
 
|-
 
|-
| [https://mega.nz/file/rUh0WCQI#rLy6foaaeagDxPKnVCVeMghO-VZr6AtTulrXmZMT6j4 1x-RoQ_nRoll]
+
| [https://github.com/Venomalia/HDcube/tree/main/v1#hdcube 4x_HDCube]
| |[[User:Kim2091|Kim2091]]
+
| Venomalia
| 1x
+
| 4x
| CC BY-NC-SA 4.0
+
| CC0-1.0
 
| ESRGAN
 
| ESRGAN
| Compression
+
| Gamecube and Wii textures.
| 2021-11-21
+
| 2022-05-10
| This model decompresses images and video compressed using RoQ. Config and presets will be added when Mega decides to let me use their site!
+
| Gamecube and Wii textures (mainly DXT and 8bit color compression). Is good for preserving fine details without affecting the original style too much, it is not suitable for pixel art, small icons and text under 16 pixel.
| none
+
| 4x_NMKD Siax
| https://imgsli.com/ODI5Njg
+
| https://imgsli.com/MTA3NDQ3
https://cdn.discordapp.com/attachments/911838608879677440/912086092755374200/unknown.png
 
https://cdn.discordapp.com/attachments/911838608879677440/912084886792314880/unknown.png
 
https://cdn.discordapp.com/attachments/911838608879677440/912085213859954788/unknown.png
 
https://cdn.discordapp.com/attachments/911838608879677440/912087766366564372/unknown.png
 
|-
 
  
|-
+
https://slow.pics/c/rmQsQ6yb
| [https://mega.nz/file/qAY11ZBR#zgLbwfSI2MDYHb0HxehLkq4cppEg-U9bplxHu5hVrgU 1x_DeRoqBeta-lite]
 
| |[[User:Kim2091|Kim2091]]
 
| 1x
 
| CC BY-NC-SA 4.0
 
| ESRGAN
 
| Compression
 
| 2022-07-26
 
| Incomplete lite model to remove ROQ compression
 
|
 
| https://cdn.discordapp.com/attachments/579685650824036387/1001671437293207672/1658887461.0522466.png
 
 
|-
 
|-
  
| [https://icedrive.net/1/43GNBihZyi NMKD h264Texturize]
+
| [https://u.pcloud.link/publink/show?code=XZTAmsXZzEO8tMJaYAF1MrKKnmOCKu7Ghedy 1x_DXTless_SourceEngine_170000_G]
| |[[User:nmkd|Nmkd]]
+
| Xeller
 
| 1x
 
| 1x
 
| WTFPL
 
| WTFPL
 
| ESRGAN
 
| ESRGAN
| Texturizing
+
| DTX5 Compression
| 2020-10-20
+
| 2021-04-10
| Tries to reverse heavy h264 compression. Fails. Can be used to texturize images though.
+
| This model is made for Source Engine textures. It tries to remove compression artifacts such as blockyness, discoloration, green tint. It does pretty well on a lot of things, realistic stuff as well, but it was mostly made to work on TF2 textures. It's made to keep as much detail as possible, without any unnecessary denoising/sharpening. Huge thanks to Twittman for assisting me along this journey.
| 4x ESRGAN
+
| 1x_Saiyajin_DeJpeg_300000_G
|  
+
| https://imgsli.com/NDk1NTk
 +
 
 +
https://imgsli.com/NDk1NjA
 +
 
 +
https://imgsli.com/NDk1NjE
 +
 
 +
https://imgsli.com/NDk1NjI
 
|-
 
|-
  
 
+
| [https://drive.google.com/file/d/1HIBRKFs7s-XhpN1p7rwAWcMTHwFRqn3n/view Trixie]
| [https://drive.google.com/drive/folders/1ZcMD3hy1YCm-hsrT6xYGvhboxSzhrJKg 1x_Filmify4K_v2_325000_G.pth]
+
| |[[User:LyonHrt|LyonHrt]]
| Muf
+
| 4x
| 1x
+
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
| Public Domain
 
 
| ESRGAN
 
| ESRGAN
| artifact
+
| Faces / Game Textures
| 2021-07-19
+
| not on discord
| This model attempts to make films upscaled to 4K with Topaz Gaia-HQ look more natural and filmic. It sharpens, adds film grain, and smooths out small artefacts from the upscaling process. I recommend adding a tiny amount of grain to the input to seed the model (you can do this in VEAI), otherwise the film grain will remain static across frames that don't move much. Pretrain model used with permission to relicense from Twittman.
+
| 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
| 1x_UnResize_V3_110000_G.pth
+
| none
| https://imgsli.com/NjE5MTE https://imgsli.com/NjE5MTI https://imgsli.com/NjE5MTM https://imgsli.com/NjE5MTQ
+
|  
 
|-
 
|-
  
|-
+
| [https://e.pcloud.link/publink/show?code=XZGmsZo1yNrot19O4R5SrPVD1GWpRy1MKX Skyrim Armory]
| [https://mega.nz/file/zAIVEaaY#AWHaCZVeIVQ8FiqsDwxQjzhiydSntFVEyGl1cCLxRzg DeIndeo (mirror)] [https://drive.google.com/file/d/1-lfgZ_6vtS9DsWU-57ig8UJExH26XmCv/view?usp=sharing DeIndeo]
+
| |[[User:Alsa|Alsa]]
| Wild West Quest#7975
 
 
| 4x
 
| 4x
| WTFPL
+
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 
| ESRGAN
 
| ESRGAN
| Indeo Compression Artifacts
+
| Game textures/equipment
 +
| not on discord
 
|  
 
|  
| Dataset: Custom
+
| Manga109Attempt
| [https://mega.nz/#!uhAVEaAA!qfwIQ44Ba3rXMAb2-M3aM3zFBCwzmyQ64IO_0O-csJE 4xESRGAN]
 
 
|  
 
|  
 
|-
 
|-
  
|}
 
  
===VHS Tapes===
+
| [https://drive.google.com/file/d/1RjVWNUFtVyz4ykqu6Pm9wmO03pPEBGqg/view Skyrim Misc]
{| class="wikitable sortable"
+
| Deorder
 +
| 4x
 +
| CC0
 +
| ESRGAN
 +
| Game textures
 +
| not on discord
 +
| Skyrim Diffuse Textures
 +
| ?
 +
|  
 
|-
 
|-
! Model Name
+
 
! Author
+
| [https://drive.google.com/file/d/1RjVWNUFtVyz4ykqu6Pm9wmO03pPEBGqg/view Skyrim Wood]
! Scale
+
| Laeris
! License
+
| 4x
! Architecture
+
|  
! Purpose (short)
 
! Date Posted
 
! Purpose (Full)
 
! Pretrained_Model_G
 
! sample
 
 
 
|-
 
| [https://icedrive.net/0/d3bwO9Byjo 1xBaldrickVHSFix_180000_G_V0.2]
 
| NimRodZorg
 
| 1x
 
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 
 
| ESRGAN
 
| ESRGAN
| VHS
+
| Game textures
| 2021-03-03
+
| not on discord
| Fixing minor VHS Chroma and Pattern Noise - NOTE: only works on deinterlaced sources
+
| Wood
| 1xESRGAN
+
| ?
 
|  
 
|  
 
|-
 
|-
  
 
|-
 
|-
| [https://drive.google.com/file/d/1w88lL_XGrE8n9sSIq97nLcVASqQNCtP6/view?usp=sharing ToonVHS]
+
| [https://mega.nz/folder/fcp2SLKC#tF3vG7zjvfpnXK7OLUaFRA 4x-SkyrimTexV2.1 and V2_Fabric]
| Redslam
+
| [[User:Kim2091|Kim2091]]
| 1x
+
| 4x
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
+
| CC BY-NC-SA 4.0
 
| ESRGAN
 
| ESRGAN
| VHS
+
| Game textures
| 2022-02-06
+
| 2021-09-25
| Best when used on cartoons, it can work on anime. Due to the dataset it does struggle a bit with orange colors and grainy dark spots. This model is meant to be used to clean up the image before using it on a 2x or 4x model.
+
| This model set upscales most Skyrim textures. I hope it helps 🙂 The base model (2.1) works well on stone, wood, metals, and most other textures. The Fabric model is a 50/50 interpolation with a stronger iteration of this model and my Fabric model. As you can tell by the name, it's intended for Fabric textures. If you want to upscale Alpha textures, use a model dedicated to it. There's an INFO file in the folder, it just explains the extra models
| 1xESRGAN
+
| 4xESRGAN
| [https://imgsli.com/OTQ0NjM/0/1 Examples of ToonVHS + other models] [https://imgsli.com/OTQ0NjQ/4/5 Examples of ToonVHS by itself]
+
| [https://cdn.discordapp.com/attachments/891300021453070387/891392169926098974/unknown.png Sample 1]  
 +
 
 +
[https://cdn.discordapp.com/attachments/891393393849147423/891397870266224731/unknown.png Sample 2]
 +
 
 +
[https://cdn.discordapp.com/attachments/891300021453070387/891395469387890738/unknown.png Sample 3]
 +
 
 +
[https://cdn.discordapp.com/attachments/891393393849147423/891398051325960202/unknown.png Sample 4]
 
|-
 
|-
  
| [https://drive.google.com/drive/folders/1NVT-5hmcrXc1d_DFjzNf-z-VfOa_xhwI?usp=sharing 2x_VHS-upscale-and-denoise_Film_477000_G]
+
| [https://drive.google.com/file/d/12fR-pWw6YL2ZbS6EDvrAeOVoQHWkWl5L/view Forest]
| Itewreed
+
| |[[User:LyonHrt|LyonHrt]]
| 2x
+
| 4x
 
|[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]
 
| ESRGAN
 
| ESRGAN
| VHS
+
| Game textures
| 2021-03-28
+
| not on discord
| VHS captures of Film material, but may work on VHS recorded native SD-TV material as well. Also useable for cleaned up source material
+
| Wood / Leaves
 
| none
 
| none
 
|  
 
|  
 
|-
 
|-
  
| [https://mega.nz/file/Z5IXBSrA#bE9VasE60zLqZ_NggOhs40Z9l2z8DapROZb5VX0F3h0 VHS-Sharpen-1x]
+
| [https://mega.nz/folder/tFhl0L5b#CK4WLHYxjX7KFXU1PLNDuQ Morrowind Mixed]
| RTX 2080 ti hoarder
+
| mkultra
| 1x
+
| 4x
 
|[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]
 
| ESRGAN
 
| ESRGAN
| VHS
+
| Game textures
| 2021-03-19
+
| not on discord
| Make old VHS footage crispy. This model will not work on video and images with noticeable JPEG/Video compression artifacts, noticeable interlacing or haloing, heavy tape distortion/artifacts and scenes with tons of detail. For best results, use a downscaled HD capture of the VHS tape you intend to use it on.
+
| Morrowind Mod Textures
| 1xESRGAN
+
| 4xESRGAN
| https://imgsli.com/NDUxNDY/2/3
+
|  
 
|-
 
|-
  
|}
+
| [https://mega.nz/folder/dVQxxITT#Bsloq9G_Xh5tsDMKDQuTTA Morrowind 2.0]
 +
| mkultra
 +
| 4x
 +
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 +
| ESRGAN
 +
| Game textures
 +
| not on discord
 +
| Morrowind Mod Textures
 +
| 4xESRGAN
 +
|
 +
|-
  
=Model Collections=
+
| [https://drive.google.com/file/d/1xPzqzTG5L6y5I0U3W0oQjexwGF5zfG32/view Map]
{| class="wikitable sortable"
+
| |[[User:LyonHrt|LyonHrt]]
 +
| 4x
 +
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 +
| ESRGAN
 +
| Game textures
 +
| not on discord
 +
| Map / Old Paper with text
 +
| none
 +
|
 +
|-
 +
 
 +
 
 +
|}
 +
 
 +
==Video Restoration==
 +
===Video Compression===
 +
 
 +
{| class="wikitable sortable"
 
|-
 
|-
 
! Model Name
 
! Model Name
Line 4,066: Line 4,281:
 
! License
 
! License
 
! Architecture
 
! Architecture
! Purpose
+
! Purpose (short)
 
! Date Posted
 
! Date Posted
! Description
+
! Purpose (Full)
! sample
+
! Pretrained_Model_G
 +
! Sample
 +
 
 
|-
 
|-
 
+
| [https://de-next.owncube.com/index.php/s/wKjLmYsq7M5JAmx 1x_cinepak_200000]
| [https://drive.google.com/drive/folders/1s39goxbCkZIk3KsM1-dJhwZyIx-iQHRy?usp=sharing OptimusPrimal's Collection]
+
| |[[User:Twittman|Twittman]]
| OptimusPrimal
+
| 1x
| Various
+
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0] (?)
| WTFPL
 
 
| ESRGAN
 
| ESRGAN
| Collection
+
| Compression
| 2022-02-28
+
| 2019-07-03
| I'm dumping all my models here. (I moved my favorites to a favorite folder). Most of these models work well on DVD resolution animation sources. These are all ESRGAN models. Everything is with the WTFPL license. So do whatever you want with them. Credit is not needed, but would be nice if you use them to make something. There's a few that are interpolations with other user's publicly available models, or interpolations of other user's publicly available models (so technically, these are not mine), so credit goes out to them, the only ones I can remember off the top of my head is @twittman with the Fatality model series, which are great, and @cd VSGAN && poetry install with the Sol Levante/American Dad2 models, and if you do 50/50 with those 2 models, you get a really good model that I called Soladad.
+
| Removal of Compression such as Cinepak, msvideo1 and Roq
 +
| none
 
|  
 
|  
 
|-
 
|-
  
| [https://mega.nz/folder/WEwUCDSJ#b1eXDT9b7yMKVlOURbR4FQ Kim's Collection]
+
|-
| Kim2091
+
| [https://de-next.owncube.com/index.php/s/wKjLmYsq7M5JAmx 1x_cinepak_alt.pth]
| Various
+
| |buildist
| CC BY-NC-SA 4.0
+
| 1x
| Various (ESRGAN and ESRGAN-lite)
+
|[https://www.gnu.org/licenses/gpl-3.0.en.html GNU GPLv3]
| Collection
+
| ESRGAN
| 2022-02-28
+
| Compression
| My (Kim's) main folder for hosting models. Every model of mine on this database links to subsections of this folder. If a link is broken or you just want to see nearly all of the models I've trained, look here.
+
|  
 +
| Cinepak (only)
 +
| none
 
|  
 
|  
 
|-
 
|-
  
| [https://mega.nz/folder/3Jo2AAAa#4CGEwUM0dKu3kkaJa-qUIA UltraMix Collection]
+
|-
| Kim2091
+
| [https://mega.nz/folder/bYxyWaZI#YQXKPVCSPF4jzHDZUcI27A DeBink v4/v5/v6]
| Various
+
| |[[User:Kim2091|Kim2091]]
 +
| 1x
 
| CC BY-NC-SA 4.0
 
| CC BY-NC-SA 4.0
| Various (ESRGAN and ESRGAN-lite)
+
| ESRGAN
| Collection
+
| Compression
| 2022-06-28
+
| 8.6.21
| This is a mixture of models based around UltraSharp and my other available models. These are usually interpolations that have separate but very helpful uses. As an example, UltraMix_Restore is a combination of UltraSharp and UniScale_Restore, and is great for video game textures.
+
| This model removes early 2000s Bink and other compression artifacts. Works well on almost any image or video compression type.
 +
| none
 
|  
 
|  
 
|-
 
|-
  
| [https://icedrive.net/s/43GNBihZyi NMKD's Collection]
 
| NMKD
 
| Various
 
| Various
 
| Various (ESRGAN and ESRGAN-lite?)
 
| Collection
 
| 2022-02-28
 
| NMKD's main folder for hosting models. Nearly every one of NMKD's models posted on this wiki link to it.
 
|
 
 
|-
 
|-
 
+
| [https://mega.nz/folder/6cxHzIxB#IN6ZgG-S0j3LjM0x5DEqew DeBink_Lite]
| [https://drive.google.com/drive/folders/1R31KChaJiw1hbVhO4NDKHNCktzSrpJpb?usp=sharing Joey's ESRGAN Bot Collection]
+
| |[[User:Kim2091|Kim2091]]
| Various
+
| 1x
| Various
+
| CC BY-NC-SA 4.0
| Various
+
| ESRGAN-lite
| Various
+
| Compression
| Collection
+
| 2021-10-30
| 2022-08-04
+
| This model was trained on lossless video frames of Metal Arms: Glitch in the System, compressed with Bink V1 for the LR frames. It's a lot more efficient than my first DeBink model, and also has less artifacts. It's not quite as robust, but the compression is barely noticeable in videos after processing with this.
| Here is a dump of all the ESRGAN-Bot models. As it was a un-curated collection, many of the models here were never officially released, but I do not know exactly which ones. If one of your models is included here and you would like me to remove it, please let me know.  
+
| none
|
+
| [https://cdn.discordapp.com/attachments/903415274521374750/904129653374078976/XB_Demo_Mov-85600_G.mp4 Video Sample]
 +
|-
 +
 
 +
 
 
|-
 
|-
 +
| [https://mega.nz/file/rUh0WCQI#rLy6foaaeagDxPKnVCVeMghO-VZr6AtTulrXmZMT6j4 1x-RoQ_nRoll]
 +
| |[[User:Kim2091|Kim2091]]
 +
| 1x
 +
| CC BY-NC-SA 4.0
 +
| ESRGAN
 +
| Compression
 +
| 2021-11-21
 +
| This model decompresses images and video compressed using RoQ. Config and presets will be added when Mega decides to let me use their site!
 +
| none
 +
| [https://imgsli.com/ODI5Njg Sample 1]
  
|}
+
[https://cdn.discordapp.com/attachments/911838608879677440/912086092755374200/unknown.png Sample 2]
  
 +
[https://cdn.discordapp.com/attachments/911838608879677440/912084886792314880/unknown.png Sample 3]
  
==Pretrained Models==
+
[https://cdn.discordapp.com/attachments/911838608879677440/912085213859954788/unknown.png Sample 4]
Looking for official models? Look here: https://upscale.wiki/wiki/Official_Research_Models
 
  
{| class="wikitable sortable"
+
[https://cdn.discordapp.com/attachments/911838608879677440/912087766366564372/unknown.png Sample 5]
 
|-
 
|-
! Model Name
 
! Author
 
! Scale
 
! License
 
! Architecture
 
! Purpose (short)
 
! Date Posted
 
! Purpose (Full)
 
! Pretrained_Model_G
 
! sample
 
  
 
|-
 
|-
| [https://mega.nz/#!D940CAKR!-BAx7tSj1CgeopGkaMW31VuPsC3hvNBL35TEVAj4LIo 1xESRGAN]
+
| [https://mega.nz/file/qAY11ZBR#zgLbwfSI2MDYHb0HxehLkq4cppEg-U9bplxHu5hVrgU 1x_DeRoqBeta-lite]
| victorca25
+
| |[[User:Kim2091|Kim2091]]
 
| 1x
 
| 1x
| Apache License 2.0
+
| CC BY-NC-SA 4.0
 
| ESRGAN
 
| ESRGAN
| Pretrained
+
| Compression
| 2019-07-06
+
| 2022-07-26
|  
+
| Incomplete lite model to remove ROQ compression
| RRDB_ESRGAN_x4.pth
 
 
|  
 
|  
 +
| [https://cdn.discordapp.com/attachments/579685650824036387/1001671437293207672/1658887461.0522466.png Sample]
 
|-
 
|-
  
 
+
| [https://icedrive.net/1/43GNBihZyi NMKD h264Texturize]
| [https://mega.nz/#!vtgSWKQT!K7Asn2zKe4N70R2aV89KEMTKhH3aiyGAAiuQDJF09qs 2xESRGAN]
+
| |[[User:nmkd|Nmkd]]
| victorca25
+
| 1x
| 2x
+
| WTFPL
| Apache License 2.0
 
 
| ESRGAN
 
| ESRGAN
| Pretrained
+
| Texturizing
| 2019-07-06
+
| 2020-10-20
|  
+
| Tries to reverse heavy h264 compression. Fails. Can be used to texturize images though.
| RRDB_ESRGAN_x4.pth
+
| 4x ESRGAN
 
|  
 
|  
 
|-
 
|-
  
| [https://mega.nz/#!uhAVEaAA!qfwIQ44Ba3rXMAb2-M3aM3zFBCwzmyQ64IO_0O-csJE 4xESRGAN]
+
 
| xinntao
+
| [https://drive.google.com/drive/folders/1ZcMD3hy1YCm-hsrT6xYGvhboxSzhrJKg 1x_Filmify4K_v2_325000_G.pth]
| 4x
+
| Muf
| Apache License 2.0
+
| 1x
 +
| Public Domain
 
| ESRGAN
 
| ESRGAN
| Pretrained
+
| artifact
| 2019-07-06
+
| 2021-07-19
|  
+
| This model attempts to make films upscaled to 4K with Topaz Gaia-HQ look more natural and filmic. It sharpens, adds film grain, and smooths out small artefacts from the upscaling process. I recommend adding a tiny amount of grain to the input to seed the model (you can do this in VEAI), otherwise the film grain will remain static across frames that don't move much. Pretrain model used with permission to relicense from Twittman.
| RRDB_ESRGAN_x4.pth
+
| 1x_UnResize_V3_110000_G.pth
|  
+
| https://imgsli.com/NjE5MTE
|-
 
  
 +
https://imgsli.com/NjE5MTI
  
| [https://mega.nz/#!6kwQiCCS!v2uN8R44vVrlzmSqffGaCnzgogkPhhl67myJbuG45SA 8xESRGAN]
+
https://imgsli.com/NjE5MTM
| victorca25
+
 
| 8x
+
https://imgsli.com/NjE5MTQ
| Apache License 2.0
+
|-
 +
 
 +
|-
 +
| [https://mega.nz/file/zAIVEaaY#AWHaCZVeIVQ8FiqsDwxQjzhiydSntFVEyGl1cCLxRzg DeIndeo (mirror)] [https://drive.google.com/file/d/1-lfgZ_6vtS9DsWU-57ig8UJExH26XmCv/view?usp=sharing DeIndeo]
 +
| Wild West Quest#7975
 +
| 4x
 +
| WTFPL
 
| ESRGAN
 
| ESRGAN
| Pretrained
+
| Indeo Compression Artifacts
| 2019-07-06
 
 
|  
 
|  
| RRDB_ESRGAN_x4.pth
+
| Dataset: Custom
 +
| [https://mega.nz/#!uhAVEaAA!qfwIQ44Ba3rXMAb2-M3aM3zFBCwzmyQ64IO_0O-csJE 4xESRGAN]
 
|  
 
|  
 
|-
 
|-
  
 +
|}
  
| [https://mega.nz/#!btwEXAoI!nNxWI89OdQCEkcRoPEFwj6GoFN8uzZyZFaCn1wbhzKY 16xESRGAN]
+
===VHS Tapes===
| victorca25
+
{| class="wikitable sortable"
| 16x
 
| Apache License 2.0
 
| ESRGAN
 
| Pretrained
 
| 2019-07-06
 
|
 
| RRDB_ESRGAN_x4.pth
 
|  
 
 
|-
 
|-
 +
! Model Name
 +
! Author
 +
! Scale
 +
! License
 +
! Architecture
 +
! Purpose (short)
 +
! Date Posted
 +
! Purpose (Full)
 +
! Pretrained_Model_G
 +
! Sample
  
 
|-
 
|-
| [https://drive.google.com/open?id=1ldwajXL50uC7PCS63B4Wato6Dnk-svNL 1x PSNR Pretrained Model]
+
| [https://icedrive.net/0/d3bwO9Byjo 1xBaldrickVHSFix_180000_G_V0.2]
| BlueAmulet
+
| NimRodZorg
 
| 1x
 
| 1x
| Apache License 2.0 (ESRGAN's license)
+
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 
| ESRGAN
 
| ESRGAN
| Pretrained
+
| VHS
| 2020-04-20
+
| 2021-03-03
| The original RRDB_PSNR_x4.pth model converted to 1x, 2x, 8x and 16x scales, intended to be used as pretrained models for new models at those scales. These are compatible with victor's 4xESRGAN.pth conversions
+
| Fixing minor VHS Chroma and Pattern Noise - NOTE: only works on deinterlaced sources
| RRDB_PSNR_x4.pth
+
| 1xESRGAN
 
|  
 
|  
 
|-
 
|-
  
 +
|-
 +
| [https://drive.google.com/file/d/1w88lL_XGrE8n9sSIq97nLcVASqQNCtP6/view?usp=sharing ToonVHS]
 +
| Redslam
 +
| 1x
 +
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 +
| ESRGAN
 +
| VHS
 +
| 2022-02-06
 +
| Best when used on cartoons, it can work on anime. Due to the dataset it does struggle a bit with orange colors and grainy dark spots. This model is meant to be used to clean up the image before using it on a 2x or 4x model.
 +
| 1xESRGAN
 +
| [https://imgsli.com/OTQ0NjM/0/1 Examples of ToonVHS + other models]
 +
 +
[https://imgsli.com/OTQ0NjQ/4/5 Examples of ToonVHS by itself]
 +
|-
  
| [https://drive.google.com/open?id=1ldwajXL50uC7PCS63B4Wato6Dnk-svNL 2x PSNR Pretrained Model]
+
| [https://drive.google.com/drive/folders/1NVT-5hmcrXc1d_DFjzNf-z-VfOa_xhwI?usp=sharing 2x_VHS-upscale-and-denoise_Film_477000_G]
| BlueAmulet
+
| Itewreed
 
| 2x
 
| 2x
| Apache License 2.0 (ESRGAN's license)
+
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 
| ESRGAN
 
| ESRGAN
| Pretrained
+
| VHS
| 2020-04-20
+
| 2021-03-28
| The original RRDB_PSNR_x4.pth model converted to 1x, 2x, 8x and 16x scales, intended to be used as pretrained models for new models at those scales. These are compatible with victor's 4xESRGAN.pth conversions
+
| VHS captures of Film material, but may work on VHS recorded native SD-TV material as well. Also useable for cleaned up source material
| RRDB_PSNR_x4.pth
+
| none
 
|  
 
|  
 
|-
 
|-
  
 
+
| [https://mega.nz/file/Z5IXBSrA#bE9VasE60zLqZ_NggOhs40Z9l2z8DapROZb5VX0F3h0 VHS-Sharpen-1x]
| [https://drive.google.com/open?id=1ldwajXL50uC7PCS63B4Wato6Dnk-svNL 4x PSNR Pretrained Model]
+
| RTX 2080 ti hoarder
| BlueAmulet
+
| 1x
| 4x
+
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
| Apache License 2.0 (ESRGAN's license)
 
 
| ESRGAN
 
| ESRGAN
| Pretrained
+
| VHS
| 2020-04-20
+
| 2021-03-19
| The original RRDB_PSNR_x4.pth model converted to 1x, 2x, 8x and 16x scales, intended to be used as pretrained models for new models at those scales. These are compatible with victor's 4xESRGAN.pth conversions
+
| Make old VHS footage crispy. This model will not work on video and images with noticeable JPEG/Video compression artifacts, noticeable interlacing or haloing, heavy tape distortion/artifacts and scenes with tons of detail. For best results, use a downscaled HD capture of the VHS tape you intend to use it on.
| RRDB_PSNR_x4.pth
+
| 1xESRGAN
|  
+
| https://imgsli.com/NDUxNDY/2/3
 
|-
 
|-
  
 +
|}
 +
 +
=Model Collections=
 +
{| class="wikitable sortable"
 +
|-
 +
! Model Name
 +
! Author
 +
! Scale
 +
! License
 +
! Architecture
 +
! Purpose
 +
! Date Posted
 +
! Description
 +
! Sample
 +
|-
  
| [https://drive.google.com/open?id=1ldwajXL50uC7PCS63B4Wato6Dnk-svNL 8x PSNR Pretrained Model]
+
| [https://drive.google.com/drive/folders/1s39goxbCkZIk3KsM1-dJhwZyIx-iQHRy?usp=sharing OptimusPrimal's Collection]
| BlueAmulet
+
| OptimusPrimal
| 8x
+
| Various
| Apache License 2.0 (ESRGAN's license)
+
| WTFPL
 
| ESRGAN
 
| ESRGAN
| Pretrained
+
| Collection
| 2020-04-20
+
| 2022-02-28
| The original RRDB_PSNR_x4.pth model converted to 1x, 2x, 8x and 16x scales, intended to be used as pretrained models for new models at those scales. These are compatible with victor's 4xESRGAN.pth conversions
+
| I'm dumping all my models here. (I moved my favorites to a favorite folder). Most of these models work well on DVD resolution animation sources. These are all ESRGAN models. Everything is with the WTFPL license. So do whatever you want with them. Credit is not needed, but would be nice if you use them to make something. There's a few that are interpolations with other user's publicly available models, or interpolations of other user's publicly available models (so technically, these are not mine), so credit goes out to them, the only ones I can remember off the top of my head is @twittman with the Fatality model series, which are great, and @cd VSGAN && poetry install with the Sol Levante/American Dad2 models, and if you do 50/50 with those 2 models, you get a really good model that I called Soladad.  
| RRDB_PSNR_x4.pth
 
 
|  
 
|  
 
|-
 
|-
  
 
+
| [https://mega.nz/folder/WEwUCDSJ#b1eXDT9b7yMKVlOURbR4FQ Kim's Collection]
| [https://drive.google.com/open?id=1ldwajXL50uC7PCS63B4Wato6Dnk-svNL 16x PSNR Pretrained Model]
+
| Kim2091
| BlueAmulet
+
| Various
| 16x
+
| CC BY-NC-SA 4.0
| Apache License 2.0 (ESRGAN's license)
+
| Various (ESRGAN and ESRGAN-lite)
| ESRGAN
+
| Collection
| Pretrained
+
| 2022-02-28
| 2020-04-20
+
| My (Kim's) main folder for hosting models. Every model of mine on this database links to subsections of this folder. If a link is broken or you just want to see nearly all of the models I've trained, look here.
| The original RRDB_PSNR_x4.pth model converted to 1x, 2x, 8x and 16x scales, intended to be used as pretrained models for new models at those scales. These are compatible with victor's 4xESRGAN.pth conversions
 
| RRDB_PSNR_x4.pth
 
 
|  
 
|  
 
|-
 
|-
  
 +
| [https://mega.nz/folder/3Jo2AAAa#4CGEwUM0dKu3kkaJa-qUIA UltraMix Collection]
 +
| Kim2091
 +
| Various
 +
| CC BY-NC-SA 4.0
 +
| Various (ESRGAN and ESRGAN-lite)
 +
| Collection
 +
| 2022-06-28
 +
| This is a mixture of models based around UltraSharp and my other available models. These are usually interpolations that have separate but very helpful uses. As an example, UltraMix_Restore is a combination of UltraSharp and UniScale_Restore, and is great for video game textures.
 +
|
 +
|-
  
| [https://mega.nz/file/cNdCTIxK#rUNM4SIjl6UWY46hd2ZPswA484y6-W0OyZ1ZA64srzg Compact Pretrained Models]
+
| [https://icedrive.net/s/43GNBihZyi NMKD's Collection]
| Zarxrax
+
| NMKD
| 1x-4x
+
| Various
| [http://www.wtfpl.net WTFPL]
+
| Various
| Real-ESRGAN "compact"
+
| Various (ESRGAN and ESRGAN-lite?)
| Pretrained
+
| Collection
| 2022-07-31
+
| 2022-02-28
| This is a collection of pretrained models for Real-ESRGAN's Compact architecture. There are 1x, 2x, and 4x models, as well as 1x and 2x "UltraCompact" and "SubCompact" models (think of these as the equivalent to ESRGAN "lite" models). By using these are pretrains for your models, you can ensure that your models are able to be interpolated with other Compact models that were trained from these. These pretrains are compatible with most existing compact models.
+
| NMKD's main folder for hosting models. Nearly every one of NMKD's models posted on this wiki link to it.
|
 
 
|  
 
|  
 
|-
 
|-
|}
 
  
==Pretrained Discriminators==
+
| [https://mega.nz/folder/BAkVjQJK#tFPH9f334BNYqEluwBVZRg Zarxrax's Collection]
{| class="wikitable sortable"
+
| Zarxrax
 +
| Various
 +
| Various
 +
| Compact
 +
| Collection
 +
| 2023-01-20
 +
| All of my compact models, including a few variations which might not be linked from elsewhere.
 +
|
 +
|-
 +
 
 +
| [https://drive.google.com/drive/folders/1R31KChaJiw1hbVhO4NDKHNCktzSrpJpb?usp=sharing Joey's ESRGAN Bot Collection]
 +
| Various
 +
| Various
 +
| Various
 +
| Various
 +
| Collection
 +
| 2022-08-04
 +
| Here is a dump of all the ESRGAN-Bot models. As it was a un-curated collection, many of the models here were never officially released, but I do not know exactly which ones. If one of your models is included here and you would like me to remove it, please let me know.
 +
|
 +
|-
 +
 
 +
|}
 +
 
 +
 
 +
==Pretrained Models==
 +
Looking for official models? Look here: https://upscale.wiki/wiki/Official_Research_Models
 +
 
 +
{| class="wikitable sortable"
 
|-
 
|-
 
! Model Name
 
! Model Name
Line 4,301: Line 4,585:
 
! Purpose (Full)
 
! Purpose (Full)
 
! Pretrained_Model_G
 
! Pretrained_Model_G
! sample
+
! Sample
  
 
|-
 
|-
| [https://u.pcloud.link/publink/show?code=kZkLGpXZLPOptIrA1fjXvFdIS0NpdzuWSfkk 1x Pretrained Discriminator Pack]
+
| [https://mega.nz/#!D940CAKR!-BAx7tSj1CgeopGkaMW31VuPsC3hvNBL35TEVAj4LIo 1xESRGAN]
| Joey
+
| victorca25
 
| 1x
 
| 1x
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
+
| Apache License 2.0
| VGG
+
| ESRGAN
| Pretrained discriminators
+
| Pretrained
| 2020-10-26
+
| 2019-07-06
| Most of these are my spongebob dataset so they will be more useful for cartoons but I did include my original faithful model discriminator.
+
|  
| none
+
| RRDB_ESRGAN_x4.pth
 
|  
 
|  
 
|-
 
|-
  
  
| [https://u.pcloud.link/publink/show?code=kZkLGpXZLPOptIrA1fjXvFdIS0NpdzuWSfkk 2x Pretrained Discriminator Pack]
+
| [https://mega.nz/#!vtgSWKQT!K7Asn2zKe4N70R2aV89KEMTKhH3aiyGAAiuQDJF09qs 2xESRGAN]
| Joey
+
| victorca25
 
| 2x
 
| 2x
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
+
| Apache License 2.0
| VGG
+
| ESRGAN
| Pretrained discriminators
+
| Pretrained
| 2020-10-26
+
| 2019-07-06
| Most of these are my spongebob dataset so they will be more useful for cartoons but I did include my original faithful model discriminator.
+
|  
| none
+
| RRDB_ESRGAN_x4.pth
 
|  
 
|  
 
|-
 
|-
  
 
+
| [https://mega.nz/#!uhAVEaAA!qfwIQ44Ba3rXMAb2-M3aM3zFBCwzmyQ64IO_0O-csJE 4xESRGAN]
| [https://u.pcloud.link/publink/show?code=kZkLGpXZLPOptIrA1fjXvFdIS0NpdzuWSfkk 4x Pretrained Discriminator Pack]
+
| xinntao
| Joey
 
 
| 4x
 
| 4x
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
+
| Apache License 2.0
| VGG
+
| ESRGAN
| Pretrained discriminators
+
| Pretrained
| 2020-10-26
+
| 2019-07-06
| Most of these are my spongebob dataset so they will be more useful for cartoons but I did include my original faithful model discriminator.
 
| none
 
 
|  
 
|  
|-
+
| RRDB_ESRGAN_x4.pth
| [https://u.pcloud.link/publink/show?code=kZkLGpXZLPOptIrA1fjXvFdIS0NpdzuWSfkk 8x Pretrained Discriminator Pack]
 
| Joey
 
| 8x
 
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 
| VGG
 
| Pretrained discriminators
 
| 2020-10-26
 
| Most of these are my spongebob dataset so they will be more useful for cartoons but I did include my original faithful model discriminator.
 
| none
 
 
|  
 
|  
 
|-
 
|-
  
| [https://u.pcloud.link/publink/show?code=kZPxrYXZOBnSmRzYywHg9heV55yt0JvKC3gk 4x_RRDB-G_ResNet-D (Both G and D)]
+
 
| Joey
+
| [https://mega.nz/#!6kwQiCCS!v2uN8R44vVrlzmSqffGaCnzgogkPhhl67myJbuG45SA 8xESRGAN]
| 4x
+
| victorca25
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
+
| 8x
| ESRGAN (G) / ResNet (D)
+
| Apache License 2.0
 +
| ESRGAN
 
| Pretrained
 
| Pretrained
| 2021-01-17
+
| 2019-07-06
| Clean bicubic downscales / Pretrained models (G/D)
+
|  
| RRDB_ESRGAN_x4_old_arch.pth
+
| RRDB_ESRGAN_x4.pth
 
|  
 
|  
 
|-
 
|-
|}
 
  
=Real-ESRGAN Compact Custom Models=
 
These are the new, small and fast Real-ESRGAN "compact" models. These are different from normal ESRGAN or Real-ESRGAN models. They can be used with chaiNNer.
 
  
{| class="wikitable sortable"
+
| [https://mega.nz/#!btwEXAoI!nNxWI89OdQCEkcRoPEFwj6GoFN8uzZyZFaCn1wbhzKY 16xESRGAN]
|-
+
| victorca25
! Model Name
+
| 16x
! Author
+
| Apache License 2.0
! Scale
+
| ESRGAN
! License
 
! Architecture
 
! Purpose (short)
 
! Date Posted
 
! Purpose (Full)
 
! Pretrained_Model_G
 
! sample
 
|-
 
 
 
| [https://mega.nz/file/cNdCTIxK#rUNM4SIjl6UWY46hd2ZPswA484y6-W0OyZ1ZA64srzg Compact Pretrained Models]
 
| Zarxrax
 
| 1x-4x
 
| [http://www.wtfpl.net WTFPL]
 
| Real-ESRGAN "compact"
 
 
| Pretrained
 
| Pretrained
| 2022-07-31
+
| 2019-07-06
| This is a collection of pretrained models for Real-ESRGAN's Compact architecture. There are 1x, 2x, and 4x models, as well as 1x and 2x "UltraCompact" and "SubCompact" models (think of these as the equivalent to ESRGAN "lite" models). By using these are pretrains for your models, you can ensure that your models are able to be interpolated with other Compact models that were trained from these. These pretrains are compatible with most existing compact models.
 
 
|  
 
|  
 +
| RRDB_ESRGAN_x4.pth
 
|  
 
|  
 
|-
 
|-
  
| [https://e1.pcloud.link/publink/show?code=kZufYRZj1INMisvFehhfS763L4ow5YUy0VV sudo_UltraCompact_2x_1.121.175_G.pth (pcloud)] / [https://www.mediafire.com/folder/7e7752gf42eky/sudo_UltraCompact_2x_1.121.175 sudo_UltraCompact_2x_1.121.175_G.pth (mediafire)]
 
| sudo
 
| 2x
 
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 
| Compact
 
| Realtime animation restauration and doing stuff like deblur and compression artefact removal
 
| 2022-05-29
 
| My first attempt to make a REALTIME 2x upscaling model while also applying teacher student learning. It beats Anime4k in every way. These benchmarks use a 3060ti and it shows that everything better than a 3060ti should be able to handle 1080p input if you create engine files and use my TensorRT code.  You can see in the readme how to convert onnx files into engines. The 2 right bars compare normal Compact2 and Ultracompact in speed, the 2 on the left showcase older apis I used which isn't too important for this showcase.  To use this, you need to use my code which is https://github.com/styler00dollar/VSGAN-tensorrt-docker. If you use Manjaro, it is also possible to pipe the data stream directly into mpv, so you can watch it in a video player without rendering a video. Yeah the model does seem a little noisy if you zoom in a lot, but don't forget that the model itself is only 1.2mb. I think it does quite well. I still try to improve on fast models, but this is good enough to share as a first model. Plz don't steal without credits, k thx.
 
| RealESRGANv2-animevideo-xsx2.pth (Teacher: RealESRGANv2-animevideo-xsx2.pth)
 
| https://cdn.discordapp.com/attachments/579685650824036387/980550564725260358/compare.png https://cdn.discordapp.com/attachments/579685650824036387/980550565354410066/unknown.png https://cdn.discordapp.com/attachments/579685650824036387/980550565723529226/compare2.webp
 
 
|-
 
|-
 +
| [https://drive.google.com/open?id=1ldwajXL50uC7PCS63B4Wato6Dnk-svNL 1x PSNR Pretrained Model]
 +
| BlueAmulet
 +
| 1x
 +
| Apache License 2.0 (ESRGAN's license)
 +
| ESRGAN
 +
| Pretrained
 +
| 2020-04-20
 +
| The original RRDB_PSNR_x4.pth model converted to 1x, 2x, 8x and 16x scales, intended to be used as pretrained models for new models at those scales. These are compatible with victor's 4xESRGAN.pth conversions
 +
| RRDB_PSNR_x4.pth
 +
|
 +
|-
 +
  
| [https://mega.nz/file/lVUDRQYa#LVyZdRS1c1Vo7IQcDA9ivBplLqPYPWjl39px7mBKuFo 2x_DigitalFlim_SubCompact]
+
| [https://drive.google.com/open?id=1ldwajXL50uC7PCS63B4Wato6Dnk-svNL 2x PSNR Pretrained Model]
| Zarxrax
+
| BlueAmulet
 
| 2x
 
| 2x
|[http://www.wtfpl.net WTFPL]
+
| Apache License 2.0 (ESRGAN's license)
| Compact
+
| ESRGAN
[nf:24 nc:8]
+
| Pretrained
| Animation
+
| 2020-04-20
| 2022-05-14
+
| The original RRDB_PSNR_x4.pth model converted to 1x, 2x, 8x and 16x scales, intended to be used as pretrained models for new models at those scales. These are compatible with victor's 4xESRGAN.pth conversions
| This was trained on the dataset that OptimusPrimal used for his DigitalFilm models. This model cleans up the image removing some noise while upscaling. This model is very fast, running around 200x the speed of a standard ESRGAN model on my system. This is primarily a proof of concept for how fast Real-ESRGAN models can get while still producing nice results.
+
| RRDB_PSNR_x4.pth
 
|  
 
|  
| https://imgsli.com/MTA3ODYx
+
|-
  
https://imgsli.com/MTA3ODY0
 
  
https://imgsli.com/MTA3ODY2
+
| [https://drive.google.com/open?id=1ldwajXL50uC7PCS63B4Wato6Dnk-svNL 4x PSNR Pretrained Model]
 +
| BlueAmulet
 +
| 4x
 +
| Apache License 2.0 (ESRGAN's license)
 +
| ESRGAN
 +
| Pretrained
 +
| 2020-04-20
 +
| The original RRDB_PSNR_x4.pth model converted to 1x, 2x, 8x and 16x scales, intended to be used as pretrained models for new models at those scales. These are compatible with victor's 4xESRGAN.pth conversions
 +
| RRDB_PSNR_x4.pth
 +
|
 +
|-
  
https://imgsli.com/MTA3ODY3
 
|-
 
  
| [https://mega.nz/file/MZdEja6I#ao8L6INGvzns4IlDpyzjpuzR-kIqm_VrC3Q8pUTZncM 1x_HurrDeblur_SubCompact]
+
| [https://drive.google.com/open?id=1ldwajXL50uC7PCS63B4Wato6Dnk-svNL 8x PSNR Pretrained Model]
| Zarxrax
+
| BlueAmulet
| 1x
+
| 8x
|[http://www.wtfpl.net WTFPL]
+
| Apache License 2.0 (ESRGAN's license)
| Compact
+
| ESRGAN
[nf:24 nc:8]
+
| Pretrained
| Animation
+
| 2020-04-20
| 2022-06-04
+
| The original RRDB_PSNR_x4.pth model converted to 1x, 2x, 8x and 16x scales, intended to be used as pretrained models for new models at those scales. These are compatible with victor's 4xESRGAN.pth conversions
| This is a sharpening/deblurring model for anime video. It was created with three goals in mind:
+
| RRDB_PSNR_x4.pth
* Be blazing fast
 
* Try to avoid enhancing noise and compression artifacts
 
* Try to avoid over-enhancing intentionally blurred parts of the image
 
Despite that last point, this is not intended to be used on modern anime which makes heavy of use depth-of-field effects.
 
 
|  
 
|  
| https://imgsli.com/MTEwOTg4
+
|-
  
https://imgsli.com/MTEwOTY2
 
  
https://imgsli.com/MTEwOTY5
+
| [https://drive.google.com/open?id=1ldwajXL50uC7PCS63B4Wato6Dnk-svNL 16x PSNR Pretrained Model]
 +
| BlueAmulet
 +
| 16x
 +
| Apache License 2.0 (ESRGAN's license)
 +
| ESRGAN
 +
| Pretrained
 +
| 2020-04-20
 +
| The original RRDB_PSNR_x4.pth model converted to 1x, 2x, 8x and 16x scales, intended to be used as pretrained models for new models at those scales. These are compatible with victor's 4xESRGAN.pth conversions
 +
| RRDB_PSNR_x4.pth
 +
|
 +
|-
  
https://imgsli.com/MTEwOTcx
 
|-
 
  
| [https://mega.nz/file/9INFwYLD#KlV11g5yDG6fWWdZDXhztnsqhuLSIUP7-Upq3szgMCQ 1x_AnimeUndeint_Compact]
+
| [https://mega.nz/file/Bc0SmIyK#ThTyculaFvrLGMIY8FicUUWBem8O_ZF8LgpPXCn7aVQ Compact Pretrained Models]
 
| Zarxrax
 
| Zarxrax
| 1x
+
| 1x-4x
|[http://www.wtfpl.net WTFPL]
+
| [http://www.wtfpl.net WTFPL]
| Compact
+
| Real-ESRGAN "compact"
| Animation
+
| Pretrained
| 2022-06-05
+
| 2022-07-31
| This model corrects jagged lines on animation that has been deinterlaced. It handles simple line doubling, line interpolation, and even Yadif-style artifacts. It can also handle sources that were resized after deinterlacing, for example resizing from ntsc to pal resolutions. If a source has been upscaled after deinterlacing, it will need to be downsized before applying this model.
+
| This is a collection of pretrained models for Real-ESRGAN's Compact architecture. There are 1x, 2x, and 4x models, as well as 1x and 2x "UltraCompact" and "SuperUltraCompact" models (think of these as the equivalent to ESRGAN "lite" models). By using these are pretrains for your models, you can ensure that your models are able to be interpolated with other Compact models that were trained from these. These pretrains are compatible with most existing compact models.
 +
|
 
|  
 
|  
| https://imgsli.com/MTExMTE0
+
|-
 +
|}
  
https://imgsli.com/MTExMTE1
+
==Pretrained Discriminators==
 
+
{| class="wikitable sortable"
https://imgsli.com/MTExMTE2
+
|-
 +
! Model Name
 +
! Author
 +
! Scale
 +
! License
 +
! Architecture
 +
! Purpose (short)
 +
! Date Posted
 +
! Purpose (Full)
 +
! Pretrained_Model_G
 +
! Sample
  
https://imgsli.com/MTExMTE3
 
 
|-
 
|-
 
+
| [https://u.pcloud.link/publink/show?code=kZkLGpXZLPOptIrA1fjXvFdIS0NpdzuWSfkk 1x Pretrained Discriminator Pack]
| [https://mega.nz/file/lNdTmSKD#R7GIISShx-zsUVwDHw0oosv72ifbB31GClXXV3UtLLo 1x_BleedOut_Compact]
+
| Joey
| Zarxrax
 
 
| 1x
 
| 1x
|[http://www.wtfpl.net WTFPL]
+
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
| Compact
+
| VGG
| Animation
+
| Pretrained discriminators
| 2022-07-25
+
| 2020-10-26
| This model helps repair color bleed and heavy chroma noise that may be present on some older footage, particularly that which was recorded on VHS. It also cleans up rainbows if they are present.
+
| Most of these are my spongebob dataset so they will be more useful for cartoons but I did include my original faithful model discriminator.
 +
| none
 
|  
 
|  
| https://imgsli.com/MTE4MjEz
+
|-
  
https://imgsli.com/MTE4MjE4
 
  
https://imgsli.com/MTE4MjE5
+
| [https://u.pcloud.link/publink/show?code=kZkLGpXZLPOptIrA1fjXvFdIS0NpdzuWSfkk 2x Pretrained Discriminator Pack]
 +
| Joey
 +
| 2x
 +
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
 +
| VGG
 +
| Pretrained discriminators
 +
| 2020-10-26
 +
| Most of these are my spongebob dataset so they will be more useful for cartoons but I did include my original faithful model discriminator.
 +
| none
 +
|
 
|-
 
|-
  
| [https://mega.nz/file/VQkDmSgb#dlVuJZQ__yAKdJaWA_4Nx52YU_c2qy_a2oG9_a98dO4 1x_Dotzilla_Compact]
+
 
| Zarxrax
+
| [https://u.pcloud.link/publink/show?code=kZkLGpXZLPOptIrA1fjXvFdIS0NpdzuWSfkk 4x Pretrained Discriminator Pack]
| 1x
+
| Joey
|[http://www.wtfpl.net WTFPL]
+
| 4x
| Compact
+
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
| Animation
+
| VGG
| 2022-12-09
+
| Pretrained discriminators
| Wipes out dot crawl and rainbows in animation.
+
| 2020-10-26
| 1x_Compact_Pretrain.pth
+
| Most of these are my spongebob dataset so they will be more useful for cartoons but I did include my original faithful model discriminator.
| https://imgsli.com/MTM4ODkz
+
| none
 +
|  
 
|-
 
|-
 
+
| [https://u.pcloud.link/publink/show?code=kZkLGpXZLPOptIrA1fjXvFdIS0NpdzuWSfkk 8x Pretrained Discriminator Pack]
| [https://mega.nz/folder/XZgSmAoa#KNKbXZVBDq4UD1NJLHm1oQ 2x / 4x-anifilm_compact]
+
| Joey
| [[User:Kim2091|Kim2091]]
+
| 8x
| 2x and 4x
+
|[https://creativecommons.org/licenses/by-nc-sa/4.0/ CC BY-NC-SA 4.0]
| CC BY-NC-SA 4.0
+
| VGG
| Compact
+
| Pretrained discriminators
| Animation
+
| 2020-10-26
| 2022-08-02
+
| Most of these are my spongebob dataset so they will be more useful for cartoons but I did include my original faithful model discriminator.
| This model is based on a private model by @eula 5600x 3070 named 4x_eula_anifilm_v1_225k. He sent me a copy of the model, and I decided to train a compact model based on it with his permission. This model seems to fix the majority of the issues the original model had while being far faster, it's just a tiny bit softer in some images.
+
| none
 
+
|  
The dataset consists of Dragon Ball movies converted to YUV24 with @sgdisk --zap-all /dev/sda's help to reduce artifacts, then upscaled with ArtClarity and eula_anifilm. LRs are the original frames right from DVD. As a result, this model corrects some color space issues. The 2x model's HRs were downscaled by 50% with Lanczos.
 
 
 
The 2x and 4x models are pretty close in output despite being trained separately. The 2x model is a bit softer overall.
 
 
 
The models in the Real-ESRGAN Compatible folder are the original output from Real-ESRGANs training code for compatibility reasons.
 
| 4x_Compact_Pretrain.pth
 
| 2x Comparison:
 
[https://cdn.discordapp.com/attachments/903415274521374750/1004152057739087972/1659478894.4425747.png Sample 1]
 
[https://imgsli.com/MTE5MjUz Slider Comparison 1]
 
 
 
4x Comparison:
 
[https://imgsli.com/MTE5MjU0 Sample 2]
 
[https://cdn.discordapp.com/attachments/903415274521374750/1004158991544365097/1659480539.543144.png Slider Comparison 2]
 
 
|-
 
|-
  
| [https://mega.nz/file/xI8D2SCJ#oQ0h3k9lemeNJcuysJql7z55WV_NhOsKRpzpErIhcGc LD-Anime_Compact]
+
| [https://u.pcloud.link/publink/show?code=kZPxrYXZOBnSmRzYywHg9heV55yt0JvKC3gk 4x_RRDB-G_ResNet-D (Both G and D)]
| Zarxrax / Skr
+
| Joey
| 2x
+
| 4x
| [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]
| Compact
+
| ESRGAN (G) / ResNet (D)
| Animation
+
| Pretrained
| 2022-12-22
+
| 2021-01-17
| I trained Skr's great LD-Anime model on compact architecture. It upscales while fixing numerous video problems, including: noise/grain, compression artifacts, rainbows, dot crawl, halos and color bleed. This compact version may look slightly worse than Skr's original model, but runs significantly faster and also retains the correct colors better than the original model did.
+
| Clean bicubic downscales / Pretrained models (G/D)
| 2x_Compact_Pretrain.pth
+
| RRDB_ESRGAN_x4_old_arch.pth
| https://imgsli.com/MTQyMzM3/0/1
+
|  
 
|-
 
|-
 +
|}
  
|}
 
  
 
=Official Models=
 
=Official Models=
Line 4,551: Line 4,826:
 
! Purpose (Full)
 
! Purpose (Full)
 
! Pretrained_Model_G
 
! Pretrained_Model_G
! sample
+
! Sample
 
|-
 
|-
  
Line 4,612: Line 4,887:
 
| upscale IRL videos. Use not recommended by creator
 
| upscale IRL videos. Use not recommended by creator
 
| none
 
| none
| https://cdn.discordapp.com/attachments/547949806761410560/813134983236419664/100k_iter.mp4
+
| [https://cdn.discordapp.com/attachments/547949806761410560/813134983236419664/100k_iter.mp4 Sample Video]
 
|-
 
|-
  
Line 4,652: Line 4,927:
 
! Purpose (Full)
 
! Purpose (Full)
 
! Pretrained_Model_G
 
! Pretrained_Model_G
! sample
+
! Sample
 
|-
 
|-
  
Line 4,692: Line 4,967:
 
! Description
 
! Description
 
! Pretrained_Model_G
 
! Pretrained_Model_G
! sample
+
! Sample
 
|-
 
|-
  
| [https://e1.pcloud.link/publink/show?code=kZfoGRZNuvokO5THhVzVLOt7ocHkR9vDdF7 sudo_rife4_269.662_testV1_scale1.pth]
+
| <small>[https://e1.pcloud.link/publink/show?code=kZfoGRZNuvokO5THhVzVLOt7ocHkR9vDdF7 sudo_rife4_269.662_testV1_scale1.pth]</small>
 
| sudo
 
| sudo
 
| 2x
 
| 2x
Line 4,704: Line 4,979:
 
| I never really mentioned it in model-releases prior since I think not too many care about interpolation here, but I trained a rife4 model for animation a some months ago, which is better than rife4 and rife4.2 imo. Thought I should also mention it here as well. I also converted it into ncnn. (Nihuis rife ncnn models are only exported with the fastest mode and not the best quality. I exported ncnn models for the most important quality settings. Due to different export/quality settings, there are multiple models. For that reason alone, my ncnn models are much better too, since nihui only exported the fastest one.) My https://github.com/styler00dollar/VSGAN-tensorrt-docker also has the rife ncnn extention, which can use VMAF, dedup, scene detection and so on, which I would recommend. My models are in that extention as well, just select model 10, 11 or 12 and use the dev docker. That test video is done with 2x framerate, enbemble True and FastMode False, combined with scene detection and dedup stuff, tta False. Towards the best quality rife can do. Plz don't steal without credits, k thx.
 
| I never really mentioned it in model-releases prior since I think not too many care about interpolation here, but I trained a rife4 model for animation a some months ago, which is better than rife4 and rife4.2 imo. Thought I should also mention it here as well. I also converted it into ncnn. (Nihuis rife ncnn models are only exported with the fastest mode and not the best quality. I exported ncnn models for the most important quality settings. Due to different export/quality settings, there are multiple models. For that reason alone, my ncnn models are much better too, since nihui only exported the fastest one.) My https://github.com/styler00dollar/VSGAN-tensorrt-docker also has the rife ncnn extention, which can use VMAF, dedup, scene detection and so on, which I would recommend. My models are in that extention as well, just select model 10, 11 or 12 and use the dev docker. That test video is done with 2x framerate, enbemble True and FastMode False, combined with scene detection and dedup stuff, tta False. Towards the best quality rife can do. Plz don't steal without credits, k thx.
 
|
 
|
| https://cdn.discordapp.com/attachments/579685650824036387/990345004260151296/ngnl_sudorife.mp4
+
| [https://cdn.discordapp.com/attachments/579685650824036387/990345004260151296/ngnl_sudorife.mp4 Sample Video]
 
|-
 
|-
  
Line 4,721: Line 4,996:
 
! Description
 
! Description
 
! Pretrained_Model_G
 
! Pretrained_Model_G
! sample
+
! Sample
 
|-
 
|-
  
Line 4,736: Line 5,011:
 
Plz don't steal without credits, k thx.
 
Plz don't steal without credits, k thx.
 
|
 
|
| https://cdn.discordapp.com/attachments/579685650824036387/1002663424695750686/output.mp4
+
| [https://cdn.discordapp.com/attachments/579685650824036387/1002663424695750686/output.mp4 Sample Video]
 
|-
 
|-
  
Line 4,783: Line 5,058:
 
| 2021-09-14
 
| 2021-09-14
 
|  
 
|  
| My ~~first~~ (ok technically second) attempt to create a video frame interpolation model and I like how it turned out. To use it, you can either use https://gitlab.com/hubert.sontowski2007/cainapp, https://github.com/styler00dollar/Colab-CAIN or the bot in [Game Upscale] (model called rvpv1 there, just use --model rvpv1). Some demo videos in pcloud, but you need to download them. The web player seems to playback in low fps. The architecture is mainly the same to the original CAIN, but i modified the padding to be zero padding instead. ``.pt`` means JIT model, `.pth`, means normal pytorch model. Architecture file is in pcloud as well. And no, no cupscale or flowframes. Dataset: Modified Animeinterp dataset
+
| My ~~first~~ (ok technically second) attempt to create a video frame interpolation model and I like how it turned out. To use it, you can either use [https://gitlab.com/hubert.sontowski2007/cainapp cain-App], [https://github.com/styler00dollar/Colab-CAIN Colab-CAIN (no longer available)] or the bot in the [https://discord.gg/cpAUpDK Game Upscale discord] (model called rvpv1 there, just use --model rvpv1). Some demo videos in pcloud, but you need to download them. The web player seems to playback in low fps. The architecture is mainly the same to the original CAIN, but i modified the padding to be zero padding instead. ``.pt`` means JIT model, `.pth`, means normal pytorch model. Architecture file is in pcloud as well. And no, no cupscale or flowframes. Dataset: Modified Animeinterp dataset
| https://files.catbox.moe/xiq9vi.mp4
+
| [https://files.catbox.moe/xiq9vi.mp4 Sample Video]
 
|-
 
|-
  

Revision as of 23:41, 27 May 2023

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.
For Datasets to train your models, go here: https://upscale.wiki/wiki/Dataset_Database
Looking for Official Models (Such as ESRGAN, BSRGAN, or Real-ESRGAN)? Check this page: https://upscale.wiki/wiki/Official_Research_Models

ESRGAN ("old Architecture") Models

These are all models that use the "old" ESRGAN architecture. There are various GUIs available to inference/upscale with these models. The only actively maintained program is chaiNNer by Joey. The others are: IEU by Honh Cupscale by NMKD.

These programs can be used to train your models: BasicSR, the official ESRGAN repository (old arch tag), victorca's traiNNer (https://github.com/victorca25/traiNNer), or sudo's colab-traiNNer (https://github.com/styler00dollar/Colab-traiNNer/).

Image scaling and Video upscaling

Universal Models

Model Name Author Scale License Architecture Purpose (short) Date Posted Purpose (Full) Pretrained_Model_G Sample
CountryRoads EzoGaming 4x WTFPL ESRGAN Universal Upscaler 2021-08-26 Streets with dense foliage in the background. Outdoor scenes. 4xPSNR Sample 1

Sample 2

Sample 3

Sample 4

Remacri Remacri Backup Mirror Foolhardy 4x CC BY-NC-SA 4.0 ESRGAN General Upscaler 2021-04-09 A creation of BSRGAN with more details and less smoothing, made by interpolating IRL models such as Siax, Superscale, Superscale Artisoft, Pixel Perfect, etc. This was, things like skin and other details don't become mushy and blurry. none - interpolated
4x-UltraSharp Kim2091 4x CC BY-NC-SA 4.0 ESRGAN Universal Upscaler 2021-10-27 This is my best model yet! It generates lots and lots of detail and leaves a nice texture on images. It works on most images, whether compressed or not. It does work best on JPEG compression though, as that's mostly what it was trained on. It has the ability to restore highly compressed images as well! If you want a more balanced output, check out the UltraMix Collection down below. It's a bunch of interpolated models based around UltraSharp and my other models 4xESRGAN Sample 1

Sample 2

Sample 3

Sample 4

UniScale-Balanced/Strong Kim2091 4x CC BY-NC-SA 4.0 ESRGAN Universal Upscaler 2021-08-23 UniScale strikes a nice balance between sharpness and realism. This model can upscale almost anything well. It was originally intended to upscale game textures, but was expanded into a universal upscaler. Interp is these two models interpolated. 4xESRGAN https://slow.pics/c/i75q0yN1
UniScaleNR-Balanced/Strong Kim2091 4x CC BY-NC-SA 4.0 ESRGAN Universal Upscaler 2021-08-23 Version of UniScale trained with camera noise injection (NR = Noise Removal). This model removes noise from images while upscaling. 4xESRGAN https://slow.pics/c/UkCwdK13
UniScale_Restore Kim2091 4x CC BY-NC-SA 4.0 ESRGAN Universal Upscaler 2021-08-27 UniScale_Restore has strong compression removal that helps with restoring heavily compressed or noisy images. It is intended to compete with BSRGAN. Trained with BSRGAN_Resize and Combo_Noise in traiNNer. 4xESRGAN Sample 1

Sample 2

UniScaleV2_Soft/Moderate/Sharp Kim2091 4x CC BY-NC-SA 4.0 ESRGAN Universal Upscaler 2021-09-06 These models work great on game textures when interpolated 50/50 with UniScale_Restore, and work amazingly on uncompressed images. DO NOT USE FOR COMPRESSED IMAGES, use the original UniScale or UltraSharp for that. 4xESRGAN Sample 1

Sample 2

Sample 3

Sample 4

realesrgan-x4minus DinJerr 4x WTFPL Real-ESRGAN General Upscaler 2022-04-19 Basically realesrgan-x4plus without the degradation training. Supposed to help retain more details, but unfortunately due to the dataset (I think) still blurs details adjacent to other objects. realesrgan-x4plus https://imgsli.com/MTA0OTA2
4x_2C2-ESRGAN_Nomos2K_200000_G.pth Joey 4x MIT 2C2-ESRGAN General Upscaler 2022-04-20 Technically my previous experiment was the pretrained model, but for all intents and purposes this was trained from scratch. Description: Pretrained model for the new architecture modification I made. You can read more about it in the github README. Basically it makes smaller ESRGAN models that theoretically can produce the same level of quality. NOTE: THIS WILL NOT WORK IN CUPSCALE, IEU, OR CHAINNER (yet)! You have to use my fork to use it for now.
FuzzyBox BlueAmulet 4x CC BY-NC 4.0 ESRGAN General upscaler 2020-07-21 Photographs, Artwork, Textures, Anything really - Tried out a new pixel loss idea based on ensuring the HR downscaled matches the LR. Colors are pretty good as well as edges, but generated details seem slightly fuzzy hence the name. RRDB_ESRGAN_x4
Lollypop LyonHrt 4x CC BY-NC-SA 4.0 ESRGAN General upscaler 2020-09-11 A universal model, that is aimed at prerendered images, but handles realistic faces, manga, pixel art and dedithering as well. Trained using the patchgan discriminator, with cx loss, cutmixup and frequency separation, it produces good results with a slight grain due to patchgan, with some sharpening using cutmixup. 4X_esrgan.pth
UniversalUpscalerV2 Mutin_Choler 4x WTFPL ESRGAN General Upscaler 2021-04-01 General Upscaler 1xESRGAN https://imgsli.com/NDc3ODA/2/1
NMKD Siax ("CX") Nmkd 4x WTFPL ESRGAN General upscaler 2020-11-06 Universal upscaler for clean and slightly compressed images (JPEG quality 75 or better) 4xPSNR

Realistic Photos

Model Name Author Scale License Architecture Purpose (short) Date Posted Purpose (Full) Pretrained_Model_G Sample
4x_RealisticRescaler_100000_G Mutin Choler 4x WTFPL Real-ESRGAN Realistic Photos 2023-01-01 This model was made to upscale realistic low-res textures that are compressed by either JPEG or BC1. From my testing, this works rather well on realistic GameCube textures such as the ones from Shrek Extra Large and the board textures from Mario Party 4. This model could also work on some real life images, especially the ones that are taken outdoors. RealESRGAN_x4plus https://imgsli.com/MTQ0Mzk1

https://imgsli.com/MTQ0MTYy

https://imgsli.com/MTQ0MjEx

https://imgsli.com/MTQ0MTQ2

https://imgsli.com/MTQ0MTQ3

https://imgsli.com/MTQ0MTQ5

https://imgsli.com/MTQ0MTUw

https://imgsli.com/MTQ0MTUx

https://imgsli.com/MTQ0MTUz

https://imgsli.com/MTQ0MTU1

https://imgsli.com/MTQ0MTU2

https://imgsli.com/MTQ0MTU4

https://imgsli.com/MTQ0MTU5

https://imgsli.com/MTQ0MTYw

https://imgsli.com/MTQ0MTYx

https://imgsli.com/MTQ0MjEz

4x-Valar musl 4x CC0 ESRGAN+ Realistic Photos 2021-09-16 Meant as an experiment to test latest techniques implemented on traiNNer, including: AdaTarget, KernelGAN, UNet discriminator, nESRGAN+ arch, noise patches, camera noise, isotropic/anisotropic/sinc blur, frequency separation, contextual loss, mixup, clipL1 pixel loss, AdamP optimizer, etc. The config file is provided on the download link above. I encourage everybody to mirror the model, distribute and modify it in anywway you want. 4x_RRDB_ESRGAN Sample 1
4x_Box buildist 4x GNU GPLv3 ESRGAN Realistic Photos 2019-06-20 RRDB_ESRGAN_x4 replacement for stuff that's supposed to look realistic. none
Nickelback BlackScout 4x GNU GPLv3 ESRGAN Realistic Photos 2020-02-12 this model aims to improve further on what has been achieved by the regular 4xESRGAN and also 4xBox. It can upscale most pictures/photos (granted they are clean enough) without destroying as much detail as the aforementioned models. It generates less moiré like patterns and keeps details without oversharpening or blurring the image too much. 4xESRGAN
NickelbackFS BlackScout 4x GNU GPLv3 ESRGAN-FS Realistic Photos 2020-07-10 This model aims to improve further on what has been achieved by the old Nickelback which was an improvement attempt over 4xESRGAN and also 4xBox. It can upscale most pictures/photos (granted they are clean enough) without destroying as much detail as Box and basic ESRGAN. 4xESRGAN
4x NMKD Superscale Nmkd 4x WTFPL ESRGAN Clean Real-World Images 2020-07-22 Upscaling of realistic images/photos with noise and compression artifacts 4xESRGAN
Misc Alsa 4x CC BY-NC-SA 4.0 ESRGAN General Upscaler The Misc model is trained on various pictures shot by myself (Alsa), including bricks, stone, dirt, grass, plants, wood, bark, metal and a few others. Manga109Attempt


Art/Pixel Art

Model Name Author Scale License Architecture Purpose (short) Date Posted Purpose (Full) Pretrained_Model_G Sample
4x_xbrz_90k LyonHrt 4x CC BY-NC 4.0 ESRGAN Pixel Art 2019-06-05 Xbrz style pixel art upscaler 4xPSNR
Rebout LyonHrt 4x CC BY-NC-SA 4.0 ESRGAN Pixel Art/Sprites 2019-08-04 For upscaling character sprites Detoon
Rebout Blend LyonHrt 4x CC BY-NC-SA 4.0 ESRGAN Pixel Art/Sprites Dataset: Custom prepared sprites from kof 94 rebout, which with gradients blending Detoon
Arzenal (v1.1) ComputerK 8x CC BY-NC-SA 4.0 ESRGAN Pixel Art Smooth general pixel art, Minecraft textures Interpolated from nmkd's pixel art upscaling models with different interpolation settings and de-dithered versions
4x_scalenx_90k LyonHrt 4x CC BY-NC 4.0 ESRGAN Pixel Art 2019-06-15 Scalenx style pixel art upscaler 4xPSNR
4x_xbrz+dd_260k LyonHrt 4x CC BY-NC 4.0 ESRGAN Pixel Art 2019-06-17 xbrz plus dedithering style pixel-art upscaling model to wiki under specialised 4x_xbrz_90k
Lady0101_208000.pth DinJerr 4x ESRGAN Pixel Art/Paintings 2019-07-25 Upscale pixel art/paintings to digital painting style WaifuGAN_v3_30000.pth
Fatality (4x_Fatality_01.pth) twittman 4x ESRGAN Pixel Art/Sprites 2019-07-29 Upscales medium resolution Sprites, dithered or undithered, can also upscale manga/anime and gameboy camera images. 4x_Faces_04_N_180000_G.pth
Fatality MK2 (4x_Fatality_MKII_90000_G.pth) twittman 4x ESRGAN Pixel Art/Sprites Dataset: Anime, Manga and some real life A previously attempted MK2
Fatal Pixels (4x_FatalPixels_340000_G.pth) twittman 4x ESRGAN Pixel Art/Sprites Dataset: Anime, Manga Fatality_MKII_90k
Faithful 2x Joey 2x CC-BY-NC 4.0 ESRGAN Pixel Art 2019-09-08 2xESRGAN
deviantPixelHD raulsangonzalo 4x You're free to use this model for all non-commercial projects and interpolation, please give credit if featured in any projects ESRGAN Pixel Art 2019-09-09 Similar to Manga109, can be used as a general digital upscaler as well as with pixel art RRDB_PSNR_x4 Video Samples
4x_BS_Deviance systemd-resolved/BlackScout 4x GNU GPLv3 ESRGAN Art 2020-02-25 This model upscales Digital Drawings. It was trained on random drawings found on DeviantArt. Mostly Landscape and Scenery and Illustrations of Characters. It does fairly well and works on many different styles. 4xESRGAN
4xSmoothRealism Joey 4x CC BY-NC 4.0 ESRGAN Pixel Art 2020-03-04 Pixel art, rocky/grainy textures? Quantization smoothing, adding detail. 4x_RRDB_PSNR_old_arch.pth
8x_HugePeeps_v1 DinJerr 8x CC BY-NC 4.0 ESRGAN Art/People 2020-06-16 Painted humans TGHQFace8x
ArtStation1337 DinJerr 4x CC BY-NC 4.0 ESRGAN Digital Art/People 2019-08-21 Mainly for digital art, but can be used to upscale pixel art.
4x_BS_DevianceV2 systemd-resolved/BlackScout 4x GNU GPLv3 ESRGAN-FS Art 2020-07-25 upscales Digital Drawings. It was trained on random drawings found on DeviantArt. Mostly Landscape and Scenery and Illustrations of Characters. It does fairly well and works on many different styles. 4xESRGAN
4x_BS_DevianceMIP systemd-resolved/BlackScout 4x GNU GPLv3 ESRGAN-FS Art 2020-07-25 upscales Digital Drawings. It was trained on random drawings found on DeviantArt. Mostly Landscape and Scenery and Illustrations of Characters. It does fairly well and works on many different styles. 4xESRGAN
4x_Struzan Laserschwert 4x CC BY-NC-SA 4.0 ESRGAN Art 2020-10-12 Upscaling airbrush/pencil-based artwork 4xPSNR Sample Gallery
4x_PixelPerfectV4_137000_G Mutin Choler 4x WTFPL ESRGAN Pixel Art/Sprites 2020-11-16 Sprite Upscaler 4xESRGAN https://imgsli.com/MzgxMTc/1/2
BigFArt DinJerr 4x CC BY-NC 4.0 ESRGAN Pixel art 2020-15-31 Larger-scaled pixels to digital painting Face Focus
MS Unpainter FoolhardyVEVO 8x CC BY-NC-SA 4.0 ESRGAN Pixel Art 2021-03-24 Low-resolution MS Paint drawings, general pixel art, and general dithered pixel art, all kinds of pixel art A mix of: 8x_glasshopper_ArzenalV1.1, 8x_glasshopper_MS-Unpainter, 8x_NMKD-Sphax + Sphax de-dither, and 8x_NMKD-YanderePixelArt4 + Yandere De-Dither, all interpolated with adjustments using the interpolate function in Cupscale
MS Unpainter De-Dither FoolhardyVEVO 8x CC BY-NC-SA 4.0 ESRGAN Pixel Art 2021-03-24 Low-resolution MS Paint drawings, general pixel art, and general dithered pixel art, all kinds of pixel art A mix of: 8x_glasshopper_ArzenalV1.1, 8x_glasshopper_MS-Unpainter, 8x_NMKD-Sphax + Sphax de-dither, and 8x_NMKD-YanderePixelArt4 + Yandere De-Dither, all interpolated with adjustments using the interpolate function in Cupscale
UltraFArt_v3 Suite DinJerr 4x CC BY-NC-SA 4.0 ESRGAN Art 2021-05-14 Illustrations with with larger shaped features (?). 4x_UltraFArt
HugePaint DinJerr 8x CC BY-NC 4.0 ESRGAN Digital Illustrations Trained on a variety of images from ArtStation HugePeeps
NXbrz Archerpolation 4x CC BY-NC-SA 4.0 ESRGAN Pixel Art 2021-06-09 Basic pixel art upscaling, for people who want a more simpler style and lightweight pixel art upscaling model. Sample Gallery

Drawn Material

Anime and Cartoons

Model Name Author Scale License Architecture Purpose (short) Date Posted Purpose (Full) Pretrained_Model_G Sample
4x-AnimeSharp Kim2091 4x CC BY-NC-SA 4.0 ESRGAN Anime or Text 2021-12-28 Interpolation between 4x-UltraSharp and 4x-TextSharp-v0.5. Works amazingly on anime. It also upscales text, but it's far better with anime content. I rebranded this model on 2/10/22 to 4x-AnimeSharp from 4x-TextSharpV1. Text Sample

Anime Sample

Anime Sample 2

4x-AnimeSharp-lite Kim2091 4x CC BY-NC-SA 4.0 ESRGAN-lite Anime 2022-02-11 This model is a lite version of AnimeSharp. It was trained using student-teacher learning (if i'm using the term properly), where the HRs are LRs upscaled by the full size AnimeSharp ESRGAN model, and the lite model is trained on those outputs as the HR. It works best on clean or slightly blurry anime. Downscale by 50% first in almost all cases Sample 1

Sample 2

Sample 3

sudo_RealESRGAN2x_3.332.758_G.pth / sudo_RealESRGAN2x_Dropout_ 3.799.042_G.pth (pcloud)

sudo_RealESRGAN2x_3.332.758_G.pth / sudo_RealESRGAN2x_Dropout_ 3.799.042_G.pth (mediafire)

sudo 2x CC BY-NC-SA 4.0 ESRGAN (6B) Anime 2022-06-25 Tried to make the best 2x model there is for drawings. I think i archived that. And yes, it is nearly 3.8 million iterations (probably a record nobody will beat here), took me nearly half a year to train. It can happen that in one edge is a noisy pattern in edges. You can use padding/crop for that. I aimed for perceptual quality without zooming in like 400%. Since RealESRGAN is 4x, I downscaled these images with bicubic. I would recommend my VSGAN code though and just load the onnx. https://github.com/styler00dollar/VSGAN-tensorrt-docker I just wanted a good 2x model for animations, but that model can also be used for wallpapers and so on. Before I hear people complaining, the dropout model is a modified architecture. Stuff like cupscale or chaiNNer won't work with pth. Load the onnx with VSGAN or chaiNNer. I did add the model before switching to dropout though, which is normal ESRGAN pth, that one should work everywhere. I also converted everything into onnx, jit and ncnn, so pretty much everything there is. If you want to use ncnn, don't use nihuis code (that also includes cupscale), these codes don't include propper tiling in C++, which is very bad for this model. I think chaiNNer should have overlap/padding with ncnn, so use that instead if you really want ncnn. Plz don't steal without credits, k thx. Pretrained_Model_G: RealESRGAN_x4plus_anime_6B.pth /
RealESRGAN_x4plus_anime_6B.pth 
(sudo_RealESRGAN2x_ 3.332.758_G.pth)
Sample 1

Sample 2

Sample 3

Sample 4

Sample 5

sudo_UltraCompact_2x_1.121.175_G.pth (pcloud) / sudo_UltraCompact_2x_1.121.175_G.pth (mediafire) sudo 2x CC BY-NC-SA 4.0 Compact Realtime animation restauration and doing stuff like deblur and compression artefact removal 2022-05-29 My first attempt to make a REALTIME 2x upscaling model while also applying teacher student learning. It beats Anime4k in every way. These benchmarks use a 3060ti and it shows that everything better than a 3060ti should be able to handle 1080p input if you create engine files and use my TensorRT code. You can see in the readme how to convert onnx files into engines. The 2 right bars compare normal Compact2 and Ultracompact in speed, the 2 on the left showcase older apis I used which isn't too important for this showcase. To use this, you need to use my code which is https://github.com/styler00dollar/VSGAN-tensorrt-docker. If you use Manjaro, it is also possible to pipe the data stream directly into mpv, so you can watch it in a video player without rendering a video. Yeah the model does seem a little noisy if you zoom in a lot, but don't forget that the model itself is only 1.2mb. I think it does quite well. I still try to improve on fast models, but this is good enough to share as a first model. Plz don't steal without credits, k thx. RealESRGANv2-animevideo-xsx2.pth

(Teacher: RealESRGANv2-animevideo-xsx2.pth)

Sample 1

Sample 2

Sample 3

2x_Bubble_AnimeScale_Compact_v1 Bubblemint#6472 2x CC BY-NC-SA 4.0 Real-ESRGAN Compact Anime or Text 2022-11-13 This is my first model, so it's not perfect, but I wanted to see if I could train an upscaling model that didn't result in a lot of detail loss and deblurring like the current Compact upscaling models. I believe I accomplished this, but I was unable to reduce contrast shifting. The contrast shifting may cause skin tones to appear incorrect on bright frames, but it's not too bad overall! I'll list a few examples below; more can be found by clicking the Overview link on the Github release page. 4x_muy4_035_1.pth https://imgsli.com/MTM0MzMx

https://imgsli.com/MTM0MzM3

https://imgsli.com/MTM0MzQw

2x_Bubble_AnimeScale_SwinIR_Small_ v1 Bubblemint#6472 2x CC BY-NC-SA 4.0 SwinIR Small Anime or Text 2022-11-13 2x_Bubble_AnimeScale_SwinIR_Small_ v1 was trained to upscale anime frames faithfully without major contrast shifting compared to my compact model. Although much slower compared to my compact model, the results look significantly better! A few example upscales are listed below; more can be found by clicking the Overview link on the Github release page. None https://imgsli.com/MTM2MjAx

https://imgsli.com/MTM2MjAy

https://imgsli.com/MTM2MjAz

escale katoumegumi_#3231 4x WTFPL ESRGAN Anime / Visual Novel Art 2022-12-17 Third iteration of my eroge upscaling model. Discriminator: Google Drive https://slow.pics/c/fqjAhnjH
4x_eula_anifilm_v1_225k eula 4x CC BY-NC-SA 4.0 ESRGAN Anime or Text 2022-12-14 Upscaling cel animation. Trained this more than a year ago, releasing cause I've got a much better v2 and v3 now. 4xPSNR https://slow.pics/c/UrnFYuXX

https://slow.pics/c/zy1Kd841

2x / 4x-anifilm_compact Kim2091 2x and 4x CC BY-NC-SA 4.0 Compact Animation 2022-08-02 This model is based on a private model by @eula 5600x 3070 named 4x_eula_anifilm_v1_225k. He sent me a copy of the model, and I decided to train a compact model based on it with his permission. This model seems to fix the majority of the issues the original model had while being far faster, it's just a tiny bit softer in some images.

The dataset consists of Dragon Ball movies converted to YUV24 with @sgdisk --zap-all /dev/sda's help to reduce artifacts, then upscaled with ArtClarity and eula_anifilm. LRs are the original frames right from DVD. As a result, this model corrects some color space issues. The 2x model's HRs were downscaled by 50% with Lanczos.

The 2x and 4x models are pretty close in output despite being trained separately. The 2x model is a bit softer overall.

The models in the Real-ESRGAN Compatible folder are the original output from Real-ESRGANs training code for compatibility reasons.

4x_Compact_Pretrain.pth 2x Comparison:

Sample 1

Slider Comparison 1

4x Comparison:

Sample 2

Slider Comparison 2

LD-Anime_Compact Zarxrax / Skr 2x CC BY-NC-SA 4.0 Compact Animation 2022-12-22 I trained Skr's great LD-Anime model on compact architecture. It upscales while fixing numerous video problems, including: noise/grain, compression artifacts, rainbows, dot crawl, halos and color bleed. This compact version may look slightly worse than Skr's original model, but runs significantly faster and also retains the correct colors better than the original model did. 2x_Compact_Pretrain.pth https://imgsli.com/MTQyMzM3/0/1
Futsuu Anime Zarxrax 2x WTFPL Compact Animation 2023-1-18 This model upscales while doing some sharpening and line darkening. Can also clean up some minor artifacts of various types. It is intended to to be a good general purpose upscaler that will work well with most animation. 2x_Compact_Pretrain.pth https://imgsli.com/MTQ4MDM2/
2xGT-evA.pth evA-01 2x CC BY-NC-SA 4.0 Real-ESRGAN Compact Anime or Text 2022-12-14 This is my first 2xcompact model, the main purpose of it is to upscale Dragon ball GT. for upscaling videos that have grain I would recommend denoising and dehaloing it before passing it to the model for temporal stability. 2x_Compact_Pretrain.pth Sample 1

Sample 2

Sample 3

Sample 4

1xEpsilon-one-compact.pth evA-01 1x CC BY-NC-SA 4.0 Real-ESRGAN Compact Anime or Text 2022-12-03 This model is far from perfect, but it does a decent job on removing dot crawl and dehalo\deblock old anime at fast rate without removing many details. chaining it with models like 2x-anifilm-compact or 2xLD-ANIME or 2x_AnimeClassics_UltraLite_510K will give good results i think. 1x_Compact_Pretrain.pth https://imgsli.com/MTM3Mzgz/2/3

https://imgsli.com/MTM3Mzg1/2/3

https://imgsli.com/MTM3Mzg2/0/1


Sol Levante NTSC2HD Phoenix 4x UNLICENSE ESRGAN Anime/Pretrained 2020-04-08 NTSC DVD-spec encode x4 scale super-resolution for Anime Drawing style content. The dataset has a LOT of data throughout almost every frame, so it had a lot of stuff to learn. The resulting DVD-spec encode also had some blocking at times so it also learned to fight off blocking. RRDB_PSNR_x4
2x_AnimeClassics_UltraLite_510K CG1989 2x CC BY-NC-SA 4.0 ESRGAN Anime/Pretrained 2022-03-08 A 2x Ultra Lite model coming in under 8MB. Trained with over 15 sets of LRs ranging in a wide amount of issues. Handles Rainbows, Dot Crawl, MPEG/H.264 Compression, and may even assist in removing halos, and fixing blurriness in certain cases. This is my first public model for everyone. Best when used on old anime that is grainy. I can't say what anime it's best suited for as I have tried multiple series, and have found it does a good job on most all the tests. I wouldn't say use this for Western Animation, but it may work. I have done a few tests that I have shown in the upscale results, but that was chained with other models to achieve such a result. This model is meant to retain the more natural look of a series. There is a color shift on the end result, not drastic, but still noticable. I figure you should fix any color issues in post that way to give a more polished upscale. Big thanks to @SaurusX for the model name, and just helping out in general with anything. 2X_DigitalFilmV5_Lite.pth Sample 1

Sample 2

Sample 3

Sample 4

600k 650k BooruGan Tal 4x WTFPL ESRGAN Anime 2021-11-26 This model is designed to mainly upscale anime artworks. If you have issues with chroma then try the 600k iterations release. 4x_Manga109Attempt Sample 1

Sample 2

Sample 3

2x_Byousoku_5_Centimeter.pth Mystery_Bullet#0642 (Shady Adel) 2x GNU GPLv3 ESRGAN Anime/Pretrained 2020-04-08 Anime landscape upscale. Trained on Frames from BluRay of Byousoku 5 Centimeter 2xESRGAN
4x_OLDIES_290000_G_FINAL solidd93110 4x WTFPL ESRGAN Anime 2020-08-16 i made this model to upscale old anime and denoise. RRDB_PSNR_x4.pth
fidelbd_pokemodel Neo-Raws#4055 2x WTFPL ESRGAN Anime 2020-08-29 Made this model to upscale old anime that looks blurry. 2xESRGAN https://imgsli.com/MjExMjQ/
4x_OLDIES_ALTERNATIVE_FINAL.pth solidd93110 4x CC BY-NC-SA 4.0 ESRGAN Anime 2020-09-01 This model was made for my project captain tsubasa anime so i don't know if it works good for anything else. just try it ;) RRDB_PSNR_x4.pth https://imgsli.com/MjEzMDU
2x_SHARP_ANIME_V1 solidd93110 2x CC BY-NC-SA 4.0 ESRGAN Anime 2020-09-12 this model has been trained to work on lines and details - works well on animes which have fairly fine lines at the base but also the video must be progressive or deinterlaced 2xPSNR.pth https://imgsli.com/MjIwMDY
MeguUp katoumegumi_#3231 GNU GPLv3. ESRGAN Anime 2020-09-16 Upscaling of lossless (uncompressed) anime art. Interpolated custom, hence the license. SYNLA: LyonHrt (bloc97 dataset, MIT license).

Fatal Pixels, Fatality Anime: twittman (CC BY-NC-SA 4.0, acquired special permission to relicense. Screencap). Deviance: BlackScout (GNU GPLv3)

NMKD UltraYandere Nmkd 4x WTFPL ESRGAN Art/Anime 2020-10-08 Highly flexible 2D Art upscaling 4xESRGAN
NMKD UltraYandere Lite Nmkd 4x WTFPL ESRGAN Lite [nf=32 nb=12] Anime 2020-10-15 Fast Anime/Art upscaling 4x_DIV2K-Lite
2x_pokemodel_lite_100000_G Neo-Raws#4055 2x WTFPL ESRGAN Anime 2020-11-01 Upscale old anime like pokemon none https://imgsli.com/Mjc3Nzk

https://imgsli.com/Mjc3Nzg

https://imgsli.com/Mjc3ODA

https://imgsli.com/Mjc3ODE

https://imgsli.com/Mjc3ODI

2x_SHARP_ANIME_V2 solidd93110 2x CC BY-NC-SA 4.0 ESRGAN Anime 2020-12-09 2x_PSNR.pth https://imgsli.com/MzI0MDk
2x_BIGOLDIES_415000_G.pth solidd93110 2x CC BY-NC-SA 4.0 ESRGAN Anime 2020-12-09 upscaling old anime. help to denoise and find lines and dehalo 2x_PSNR.pth https://imgsli.com/MzI0MDc
NMKD YandereNeo Nmkd 2x WTFPL ESRGAN (Lite) Anime 2021-01-26 https://i.imgur.com/oxs71v5.png


NMKD YandereNeo Nmkd 4x WTFPL ESRGAN (Lite) Anime 2021-01-26 4x_DIV2K-Lite_1M https://i.imgur.com/oxs71v5.png
2x_Waifaux-NL3-SuperLite Joey 2x CC BY-NC-SA 4.0 ESRGAN 2021-02-23 Trained this model to see how it would work trying to essentially get the same results as Waifu2x but with ESRGAN none
2x_Waifaux-NL3-SRResNet Joey 2x CC BY-NC-SA 4.0 EDSR (SRResNet) 2021-02-25 Emulating Waifu2x at Noise Level 3 NOTE: You can't use this with regular esrgan forks or the bot, it has to be run through basicsr none
4x_Training4Melozard_Anime Joey 4x Whatever @B.Melozard2 wants ESRGAN Anime 2021-03-17 RRDB_ESRGAN_x4_old_arch
2x_LD-Anime_Skr_v1.0 Skr 2x CC BY-NC-SA 4.0 ESRGAN Denoise/Dehalo 2021-04-17 Denoise, dehalo, derainbow old anime 2xESRGAN https://imgsli.com/NTA3MDU/
2x_KemonoScale_v2 EzoGaming 2x CC BY-NC-SA 4.0 ESRGAN Anime 2021-06-05 Upscaling frames from Irodori anime (namely kemono friends) from 540p (the source render resolution) to 1080p, low resolution flat shaded art, de-JPEG of the aforementioned x2_CGIMaster_v1 Sample
4x_muy4_035_1.pth katoumegumi_ 4x WTFPL ESRGAN Anime 2021-07-18 Upscaling of anime art (Specifically visual novel CG art) Sample 1

(Hakurei Reimu from Touhou Project illustrated by "okawa friend")

Sample 2

(Kitaooji Karen from Making * Lovers developed by SMEE.)(edited)

Falcon Fanart LyonHrt 4x CC BY-NC-SA 4.0 ESRGAN Anime not on discord 4xPSNR
1x_BroadcastToStudioLite_485k SaurusX 1x CC BY-NC-SA 4.0 ESRGAN Animation 2022-03-05 Improvement of low-quality cartoons from broadcast sources. Will greatly increase the visual quality of bad broadcast tape sources of '80s and '90s cartoons (e.g. Garfield and Friends, Heathcliff, DuckTales, etc). Directly addresses chroma blur, dot crawl, and rainbowing. You're highly advised to take care of haloing beforehand in your favorite video editor as the model will not fix it and may make existing halos more noticeable. Sample 1

Sample 2

Sample 3

2x_SwatKats_154000_G SaurusX 2x CC BY-NC-SA 4.0 ESRGAN Animation 2021-11-17 In addition to removing the vertical blur, the model upscales, sharpens and will remove MPEG-2 artifacting and a small amount of rainbowing and dot crawl. Another series afflicted with the vertical blur is Avatar the Last Airbender, which can be repaired by this model. The video fed into the model MUST be 540 vertical for the deblur to work properly. 2xESRGAN https://imgsli.com/NzM2MDY

https://imgsli.com/ODIxMjU

1x_SwatKatsLite_360000_G.pth SaurusX 1x CC BY-NC-SA 4.0 ESRGAN Animation 2021-11-17 Fix vertical blur / split lines / shadowing. A 1x lite model of my 2xSwatKats. Resolves the same video problems as before, but 1x and faster and meant for chaining to other 2x models (or whatever). Input MUST be 540 vertical as the blur problem is very resolution sensitive. 2xESRGAN https://slow.pics/c/F4ilU8PM
1x_SheeepIt! Kim2091 1x CC BY-NC-SA 4.0 ESRGAN Animation 2022-07-21 Restoring frames from the show "Sheeep" This model was trained to restore "Sheeep" while retaining and enhancing the noise present in the show. The model amazingly finished training in only 8.8k iterations with no pretrain, with a dataset of only 4 image pairs. This model should work well for animes and cartoons with a lot of grain present. There are some slight haloing issues in dark colors unfortunately, but I was unable to fix it. Sample 1

Sample 2

Sample 3

Sample 4

2x-UniScale_CartoonRestore-lite Kim2091 2x CC BY-NC-SA 4.0 ESRGAN Animation, Pixel Art 2021-08-31 This model has VERY strong compression removal and line restructuring that allows it to restore any heavily compressed drawings, animation, cartoons, or anime. Also works on games as well as DDS compression. It renders frames very quickly and is very viable for restoring videos. 4x-MMScale Comparisons:

Sample 1

Sample 2

Sample 3

Sample 4

Sample 5

Video Sample

Original Video

2x_ATLA_KORRA_336200_G.pth aptitude 2x WTFPL ESRGAN Animation 2021-07-22 Upscaling of Animation based on The Legend of Korra. 2xESRGAN.pth https://imgsli.com/NjIzMzU

https://imgsli.com/NjIzMTY

https://imgsli.com/NjIzMTc

https://imgsli.com/NjIzMTg

https://imgsli.com/NjIzMTk

https://imgsli.com/NjIzMjA

https://imgsli.com/NjIzMjE

https://imgsli.com/NjIzMjQ

https://imgsli.com/NjIzMjc

https://imgsli.com/NjIzMjY

https://imgsli.com/NjIzMjU

8x_BoyMeBob-Redux Joey 8x CC BY-NC-SA 4.0 ESRGAN+ (Joey's fork, eFonte fork, or iNNfer required to use) Animation 2021-08-19 Upscaling cartoons 8xBoyMeBob (unreleased) which used 8xESRGAN Examples
Spongebob (4x_SpongeBob_235000_G / 4xSpongebob) Joey 4x CC-BY-NC 4.0 ESRGAN Animation 2019-09-12 4xESRGAN
Spongebob De-Quantize Joey 1x CC-BY-NC 4.0 ESRGAN Animation - Quantized 2019-11-02 Removed color quantization/indexing and dithering from cartoon style images and textures 1x_1xDEDITHER_32_512_126900_G
4x Spongebob v6 Deblur Joey 4x CC-BY-NC 4.0 ESRGAN Animation 2019-11-08 After not being entirely happy with the main Spongebob v6 model, I trained a new one with blurring OTF options and two different downscale types. This one is much better in my opinion. 4x Spongebob v6
4x Spongebob v6 De-Quantize Joey 4x CC-BY-NC 4.0 ESRGAN Animation - Quantized 2019-11-08 A model I trained to do both de-quantizing as well as upscaling. The results are pretty blurry but it works decently for what it is. 4x Spongebob v6
4x Spongebob v6 Joey 4x CC-BY-NC 4.0 ESRGAN Animation 2019-11-08 New version of the Spongebob model. Ideally it's a lot sharper and cleaner but I'm still not sure if it works better than the old one. From what I can tell it's better in many cases. 4xESRGAN
SpongeBob.CEL.2.HD.125ki.499e-PHOENiX.pth Phoenix 4x GPLv3 - SHARE YOUR CHANGES - If you use this as a pretrained model, share your new model!!! ESRGAN Animation 2019-12-24 Restorative CEL Animation MPEG-1 and 2 Model specifically crafted for SpongeBob S01. RRDB_PSNR_x4.pth
1x_SpongeBC1-Lite Joey 1x CC BY-NC-SA 4.0 BC1/DXT1 compression 2020-11-01 First ever lite BC1/DXT1 model. Probably only useful for cartoony textures like those in spongebob games or other cartoon licensed games. 1x_DIV2K-Lite_80k
4x_SpongeBob-Reloaded Joey 4x CC BY-NC-SA 4.0 ESRGAN Animation, Pixel Art 2020-12-11 4xSpongeBob
4x_SpongeBob-Reloaded-SWAG Joey 4x CC BY-NC-SA 4.0 ESRGAN Animation, Pixel Art 2020-12-11 4xSpongeBob
American Dad 2 (American.Dad.2.HD.150ki.5e-PHOENiX.pth) PHEONiX (?) 4x ESRGAN Animation American Dad
FatalimiX Twittman 4x CC BY-NC-SA 4.0 (?) ESRGAN Animation 2019-10-29 Comic and Cartoon style images 4x_Fatality_MKII_90000_G_02.pth
Comic Book LyonHrt 4x CC BY-NC-SA 4.0 ESRGAN Animation Comic / Drawings. Trained on Custom (Spider-Man) dataset none (no interpolation)
Fatal_Anime Twittman 4x CC BY-NC-SA 4.0 (?) ESRGAN Animation Trained on Anime and Manga images none (no interpolation)
4x_AbeScale mitch#1337 4x GNU ESRGAN Linework Cartoons Dataset: LR: Clone High DVD, HR: Re-illustrated Vectors from Frames 4xESRGAN
2x_Loyaldk-SuperPony_370000 V1.0 ChaseMMD 2x CC BY-NC-SA 4.0 ESRGAN Animation 2021-02-24 Upscale MLP episodes 2x_ERGAN https://imgsli.com/NDIwNTY

https://imgsli.com/NDIwNTg

https://imgsli.com/NDIwNjE

https://imgsli.com/NDIwNjI

https://imgsli.com/NDIwNjM

2x_Loyaldk-LitePony_380500 V1.0 ChaseMMD 2x CC BY-NC-SA 4.0 ESRGAN Animation 2021-03-01 Upscale MLP episodes. Liter version. 2x_ESRGAN
2x_Loyaldk-LitePony_500000_V2.0 ChaseMMD#6957 2x CC BY-NC-SA 4.0 ESRGAN (NF=32, NB=12) Animation 2021-04-06 Upscale MLP episodes. Liter version. Able to handle compression better compared to V1.0 and no longer creates halo's/rainbows. Good for Vector 2D art however converts detail to blobs. none
2x_Loyaldk-MediumPony_500000_V2.0 ChaseMMD 2x CC BY-NC-SA 4.0 ESRGAN (NF=48, NB=18) Animation 2021-04-11 Upscale MLP episodes. Liter version. Able to handle compression better compared to V1.0 and no longer creates halo's/rainbows. Good for Vector 2D art however converts detail to blobs. Leaves more blobs when converting detail to blobs compared to litepony none
4x_Loyaldk-LitePony_500000_V2.0 ChaseMMD 4x CC BY-NC-SA 4.0 ESRGAN (NF=32, NB=12) Animation 2021-04-12 Good for Vector 2D art however converts detail to blobs. Quality seems weak with this one. However posting anyways to see if a use is found. none
2x_Loyaldk-SuperPony_500000_V2.0 ChaseMMD 2x CC BY-NC-SA 4.0 ESRGAN Animation 2021-04-30 Upscale MLP episodes. Full version. Able to handle compression better compared to V1.0 and no longer creates halo's/rainbows. Good for Vector 2D art however converts detail to blobs. none
4x_Loyaldk-SuperPony_500000_V2.0 ChaseMMD 4x CC BY-NC-SA 4.0 ESRGAN Animation 2021-04-30 Upscale MLP episodes. Full version. Able to handle compression better compared to V1.0 and no longer creates halo's/rainbows. Good for Vector 2D art however converts detail to blobs. none
4x_Loyaldk-MediumPony_500000_V2.0 ChaseMMD 4x CC BY-NC-SA 4.0 ESRGAN Animation 2021-04-30 Upscale MLP episodes. Full version. Able to handle compression better compared to V1.0 and no longer creates halo's/rainbows. Good for Vector 2D art however converts detail to blobs. none
Platoon Model Set ChaseMMD 4x CC BY-NC-SA 4.0 ESRGAN Animation 2021-06-07 Upscale Anime while keeping as much of the original detail as possible. Isn't among the sharpest models and if the anime is somewhat blurry it will retain that detail but add more pixels to give a smoother edge. Doesn't change the color very much and may make some scene's very slightly darker. Pony Models
Anime and Cartoon Restoration
Model Name Author Scale License Architecture Purpose (short) Date Posted Purpose (Full) Pretrained_Model_G Sample
1x_HurrDeblur_SuperUltraCompact Zarxrax 1x WTFPL Compact

[nf:24 nc:8]

Animation 2022-06-04 This is a sharpening/deblurring model for anime video. It was created with three goals in mind:
  • Be blazing fast
  • Try to avoid enhancing noise and compression artifacts
  • Try to avoid over-enhancing intentionally blurred parts of the image

Despite that last point, this is not intended to be used on modern anime which makes heavy of use depth-of-field effects.

https://imgsli.com/MTEwOTg4

https://imgsli.com/MTEwOTY2

https://imgsli.com/MTEwOTY5

https://imgsli.com/MTEwOTcx

1x_AnimeUndeint_Compact Zarxrax 1x WTFPL Compact Animation 2022-06-05 This model corrects jagged lines on animation that has been deinterlaced. It handles simple line doubling, line interpolation, and even Yadif-style artifacts. It can also handle sources that were resized after deinterlacing, for example resizing from ntsc to pal resolutions. If a source has been upscaled after deinterlacing, it will need to be downsized before applying this model. https://imgsli.com/MTExMTE0

https://imgsli.com/MTExMTE1

https://imgsli.com/MTExMTE2

https://imgsli.com/MTExMTE3

1x_BleedOut_Compact Zarxrax 1x WTFPL Compact Animation 2022-07-25 This model helps repair color bleed and heavy chroma noise that may be present on some older footage, particularly that which was recorded on VHS. It also cleans up rainbows if they are present. https://imgsli.com/MTE4MjEz

https://imgsli.com/MTE4MjE4

https://imgsli.com/MTE4MjE5

1x_Dotzilla_Compact Zarxrax 1x WTFPL Compact Animation 2022-12-09 Wipes out dot crawl and rainbows in animation. 1x_Compact_Pretrain.pth https://imgsli.com/MTM4ODkz

Digital Animation

Model Name Author Scale License Architecture Purpose (short) Date Posted Purpose (Full) Pretrained_Model_G Sample
4x_DigitalFake-2.1 Joey 4x CC BY-NC-SA 4.0 ESRGAN Digital Animation 2021-02-18 Replica of DigitalFrames 2.1 but interpolatable RRDB_ESRGAN_x4_old_arch.pth
4X_KCJPUNK_1.0233089 G.pth KCJPUNK 4x CC BY-NC-SA 4.0 ESRGAN Digital Animation 2021-03-23 Up-scaling Digital Animation 4x_fatal_Anime_500000_G
2X_KcjpunkAnime_2.0_Lite_196496_G KCJPUNK 2x CC BY-NC-SA 4.0 ESRGAN-lite Digital Animation 2021-03-23 Up-scaling Digital Animation This is my first attempt to make Light model so I started with 2x version. This model is much faster and give better results than my previous one.
2x_DigitoonLite_216k SaurusX 2x CC BY-NC-SA 4.0 ESRGAN-lite Digital Animation 2022-07-05 Meant as a versatile model for upscaling high detail digital anime and cartoons. Has debanding, MPEG-2 correction, and halo reduction. Trained to handle both 4:3 and 16:9 DVD material with equal efficacy. Will retain a lot of textures except for the really high freq stuff. 2x_DigiGradients_Lite_486k.pth https://imgsli.com/MTE1Mzg4
2x_DigiGradients_Lite_486k SaurusX 2x CC BY-NC-SA 4.0 ESRGAN-lite Digital Animation 2022-08-24 A very focused model meant for upscaling the TMNT 2003 DVDs. Degradations were added via AVISynth in order to match the video on the TMNT 2003 DVDs to correct the source problems. Problems corrected include aliased red chroma, chroma vertical blur, bad deinterlacing, banding, compression "grain", and poor animation line detail. The AVS scripts for the LR's were run through HCEnc to get authentic low bit rate MPEG-2 artifacts for fixing. By design, the final model gives a very digital-looking result and does not do a good job of retaining textures as the style of TMNT 2003 is all flats and gradients. none https://imgsli.com/MTAxNjc1

https://imgsli.com/ODQ2NzI

2x_DigitalFlim_SuperUltraCompact Zarxrax 2x WTFPL Compact

[nf:24 nc:8]

Animation 2022-05-14 This was trained on the dataset that OptimusPrimal used for his DigitalFilm models. This model cleans up the image removing some noise while upscaling. This model is very fast, running around 200x the speed of a standard ESRGAN model on my system. This is primarily a proof of concept for how fast Real-ESRGAN models can get while still producing nice results. https://imgsli.com/MTA3ODYx

https://imgsli.com/MTA3ODY0

https://imgsli.com/MTA3ODY2

https://imgsli.com/MTA3ODY3

Manga

Model Name Author Scale License Architecture Purpose (short) Date Posted Purpose (Full) Pretrained_Model_G Sample
FSMangaV2 Jacob_ 4x CC BY-NC-SA 4.0 ESRGAN Manga 2020-12-14 Manga-style images with or without dithering - cartoons, maybe pixel art, etc RRDB_PSNR_x4 https://i.imgur.com/oVBVthF.png
4x_eula_digimanga_bw_v1_860k end user license agreement#9756 4x CC BY-NC-SA 4.0 ESRGAN Manga 2022-1-26 Black and white digital manga with halftones. 4xPSNR Clean

Heavy Degradation

4x_eula_digimanga_bw_v2_nc1_307k end user license agreement#9756 4x CC BY-NC-SA 4.0 ESRGAN Manga 2022-08-17 Vast improvement over v1 in low frequency detail; moiré and artifacting reduced significantly and less random noise from JPEG artefacts in the input. Also now only works on 1 channel images, so it runs slightly faster on average and resulting images are much smaller but it might not work on some ESRGAN implementations, I personally recommend using chaiNNer. v1 may still be better in some edge cases.

There's also a supplementary 1x model that denoises very low quality LRs and smooths halftones so the image works better with the 4x model. Only trained it to help build the dataset and it's useless for already decent-ish LRs but may help you in some situations.

4x_eula_digimanga_bw_v1_860k https://slow.pics/c/Asbu0xgz
MangaScaleV3 Bunzero++ 2x CC BY-NC-SA 4.0 ESRGAN+ Manga 2021-05-24 To upscale manga including halftones, instead of trying to smooth them out. unreleased model https://imgsli.com/NTUyMjg
Manga109Attempt Kingdomakrillic 4x CC BY 4.0 ESRGAN Anime / Manga Pretrained model: 4xPSNR Manga109


Drawings

Model Name Author Scale License Architecture Purpose (short) Date Posted Purpose (Full) Pretrained_Model_G Sample


KDM003_scans_1x.pth BoxDrop 1x ESRGAN Art 2019-07-29 Clean up model for scanned illustrations. - Made to remove moire patterns, reduce small imperfections, and correct mild compression artifacts in scanned Kingdom Death: Monster illustrations. CMYK printing often shifts colors, so this is intended to reverse that color shifting as well. Failed attempts based on ESRGAN_1x_JPEG_80to100
DigiPaint Rastrum 4x CC0 ESRGAN Art 2019-09-07 Digital Art Upscaler 4xfalcoon300(manga).pth

Cel Animation

Model Name Author Scale License Architecture Purpose (short) Date Posted Purpose (Full) Pretrained_Model_G Sample
DigitalFilmV5 Lite OptimusPrimal 2x WTFPL ESRGAN Traditional Animation 2021-02-19 Upscaling Dragon Ball Z DBox DVD's. - grainy sources. It keeps some grain, but also doing some cleaning, sharpening, and fixing. none https://imgsli.com/NDE0MzQ

https://imgsli.com/NDE0MzU

https://imgsli.com/NDE0MzY

https://imgsli.com/NDE0Mzc

https://imgsli.com/NDE0Mzg

https://imgsli.com/NDE0Mzk

4x_HellinaCel VGDCKeroro 4x fuckin do whatever ESRGAN Traditional Animation 2021-03-21 A rougher alternative to 4xCelFrames with a focus on realistic looking cels over nice looking cels. It's trained on DetoriationFrames LRs, so if you give it an image straight from the source it will go twice as hard on it. This can be used to your advantage, though. I recommend cleaning it up before running DetoriationFrames, or else it will come out rough. 4xESRGAN https://imgsli.com/NDU2MjU


Image Restoration

Image Compression

JPEG Artifacts
Model Name Author Scale License Architecture Purpose (short) Date Posted Purpose (Full) Pretrained_Model_G Sample
JPG (00-20%) BlueAmulet 1x CC BY-NC 4.0 ESRGAN JPEG Trained on Custom (CC0 Textures) JPEG models by Alsa
JPG (20-40%) BlueAmulet 1x CC BY-NC 4.0 ESRGAN JPEG Trained on Custom (CC0 Textures) JPEG models by Alsa
JPG (40-60%) BlueAmulet 1x CC BY-NC 4.0 ESRGAN JPEG Trained on Custom (CC0 Textures) JPEG models by Alsa
JPG (60-80%) BlueAmulet 1x CC BY-NC 4.0 ESRGAN JPEG Trained on Custom (CC0 Textures) JPEG models by Alsa
JPG (80-100%) BlueAmulet 1x CC BY-NC 4.0 ESRGAN JPEG Trained on Custom (CC0 Textures) JPEG models by Alsa
JPG (00-20%) Alsa 1x CC BY-NC-SA 4.0 ESRGAN JPEG Trained on Custom (Photos / Manga)
JPG (20-40%) Alsa 1x CC BY-NC-SA 4.0 ESRGAN JPEG Trained on Custom (Photos / Manga)
JPG (40-60%) Alsa 1x CC BY-NC-SA 4.0 ESRGAN JPEG Trained on Custom (Photos / Manga)
JPG (60-80%) Alsa 1x CC BY-NC-SA 4.0 ESRGAN JPEG Trained on Custom (Photos / Manga)
JPG (80-100%) Alsa 1x CC BY-NC-SA 4.0 ESRGAN JPEG Trained on Custom (Photos / Manga)
Kim2091_DeJpeg_v0 Kim2091 1x CC BY-NC-SA 4.0 ESRGAN Superlite JPEG Model I forgot to release. This doesn't totally remove JPEG artifacts, but it does a decent job at a fast rate. It seemingly does a better job of retaining detail than some other JPEG models. The model is incomplete, I need to train it further on compact rather than ESRGAN. This is just a temporary release Custom JPEG dataset Sample 1

Sample 2

4x NMKD-PatchySharp nmkd 4x WTFPL ESRGAN (Old Arch) Art/CGI 2020-09-12 Upscaler for clean images or images with compression artifacts (jpeg quality >75) - Produces very sharp lines/edges due to NN-Filtered HR images. Proven to produce very, very good results on drawings (sharp lines) and CGI, but should also work pretty well for real-world images. 4xESRGAN https://i.imgur.com/PzFkmAg.png
1x_JPEG (by compression level) BlueAmulet 1x CC BY-NC 4.0 ESRGAN JPEG 2019-08-01 A collection of models for jpeg artifact removal, examples are included in the provided link Alsa Sample Gallery
DeJpeg Fatality_PlusULTRA! Twittman 1x CC BY-NC-SA 4.0 (?) ESRGAN JPEG 2019-10-29 1x_DeJpeg_Fatality_01_200000_G.pth
DeCompress Kim2091 4x CC BY-NC-SA 4.0 ESRGAN JPEG (works on many compression formats) 2021-09-07 USE 4x-UltraSharp INSTEAD - This is UniScaleV2, but intended for images with compression. Seems to work best on realistic images. 4xESRGAN Sample 1

Sample 2

Sample 3

Sample 4

JPG PlusULTRA (1x_jpg_PlusULTRA_130000.pth) Twittman 1x CC BY-NC-SA 4.0 (?) ESRGAN JPEG 2019-10-29 1x_DeJpeg_Fatality_01_200000_G.pth
SaiyaJin DeJpeg Twittman 1x CC BY-NC-SA 4.0 (?) ESRGAN JPEG 2019-10-29 Dataset: Anime, real life, manga, generated 1x_DeJpeg_Fatality_PlusULTRA_200000_G.pth Example 1

Example 2

Example 3

1x_JPEGDestroyer BlackScout 1x GNU GPLv3 ESRGAN JPEG 2020-02-19 This model is meant to reduce or eliminate JPEG Compression without making the original images too smooth or killing detail. It manages to do a fairly good job but don't expect overly compressed images to work with this. 1xESRGAN + previous attempts
4xJaypeg90 Jacob 4x GNU GPLv3 ESRGAN JPEG compression 2020-07-20 Photos/realistic 3D with JPEG compression, quality 85-95 and 4:2:0 chroma subsampling. - Created for Myst3 images, since all have 4:2:0 chroma subsampling and existing JPEG models did not give good results. Favors smoothing over over-sharpening. 4xESRGAN https://i.imgur.com/n09m2Rt.png
1x_SBDV-DeJPEG-Lite Joey 1x CC BY-NC-SA 4.0 ESRGAN (lite) JPEG compression 2020-10-10 1x_DIV2K-Lite_80k.pth
NMKD Jaywreck3 & Jaywreck3-Soft (Lite) Nmkd 1x WTFPL ESRGAN Lite [nf=32 nb=12] JPEG compression 2020-10-13 1xESRGAN 1xESRGAN
MangaJPEG Bunzero++ 1x CC BY-NC-SA 4.0 ESRGAN JPEG 2021-06-17 Remove JPEG artifacts from manga without destroying screentone and other details none Sample
Aliasing
Model Name Author Scale License Architecture Purpose (short) Date Posted Purpose (Full) Pretrained_Model_G Sample
SS Anti Alias 9x BlueAmulet 1x CC BY-NC 4.0 ESRGAN Anti aliasing Dataset: Custom (9x Supersampling AA) none (no interpolation)
Anti Aliasing Twittman 1x CC BY-NC 4.0 ESRGAN Anti aliasing / Images with pixelated edges Dataset: Custom (?) none (no interpolation)
1x_GainRESV3 (Aggro,Natural,Passive) CF2lter 1x WTFPL ESRGAN Anti aliasing / Deblur 2022-02-26 To eliminate aliasing and general artifacts caused by not enough resolution while bringing out details Im stopping its training here because it's getting worse, i think of some aligment issues by game's rendering pipeline + downscaling... Dataset: 5K resolution shots from paladins rendered by 200% for hr and 37.5%(1080p) for lr then downscaled to 1080 BCGONE_DetailedV2 Sample 1

Sample 2

GIF
Model Name Author Scale License Architecture Purpose (short) Date Posted Purpose (Full) Pretrained_Model_G Sample
DeGif nmkd 2x WTFPL ESRGAN GIF Restoration GIF Restoration 2xESRGAN
DDS (BC1/DXT1, BC3/DXT5 Compression)
Model Name Author Scale License Architecture Purpose (short) Date Posted Purpose (Full) Pretrained_Model_G Sample
1x_BC1-smooth2.pth BlueAmulet 1x CC BY-NC 4.0 ESRGAN BC1 Compression 2019-09-05 A model to help remove compression artifacts in BC1-BC3/DXT1-DXT5 compressed images (these all have color encoded the same way) none
x1_ITF_SkinDiffDDS_v1 intheflesh#3116 1x CC BY-NC 4.0 ESRGAN BC1 Compression 2022-08-24 Removes banding, blocking, dithering, aliasing, noise and color tint on DDS Compressed Skin Diffuse Textures.

This should work extremely well on most modern DDS compression types. The training set was compressed with BC3/DXT5, BC3/DXT5 Fast, BC2/DXT3, BC2/DXT3 Fast, and a small number of JPEG compressed images to cover outliers.

This model is trained to remove the slight green color tint that DDS compression tends to add to skin textures, so the model output will not match the original color tone of the input image. This is the desired result though, as DDS compression shifts the colors to a sickly green tint and this model corrects that to more natural color tones.

The training set included faces, body parts, eyes, mouths and hair in a variety of skin types and tones so it should work well on most related diffuse textures.

However it's not just limited to skin, many other images and textures can be cleaned with this model. Designed to be used as a first step cleaning pass before applying additional models after. Check out the other ITF Models.

1x_BC1-smooth2.pth https://imgsli.com/MTIyMzE5

https://imgsli.com/MTIyMzIw

https://imgsli.com/MTIyMzIx

1x_DXTDecompressor_Source_V3-300000_G JosephtheKP 1x WTFPL ESRGAN DXT1 Compression 2021-12-21 Removing compression artifacts from DXT1 compressed textures. This model was created to remove DXT1 compression artifacts from textures imported into the Source Engine. Compressed textures in the engine sometimes have a green-tint which this model also corrects. The data for this model contained a good mix of diffuse textures and normal maps which means this model is pretty good at removing compression from normals as well. Creating this model was a real learning experience for me and I hope someone finds a good use for it. https://imgsli.com/ODcwNDg

https://imgsli.com/ODcwNDk

https://imgsli.com/ODcwNTA

https://imgsli.com/ODcwNTE

1x_DEDXT CF2lter 1x WTFPL ESRGAN DXT compression 2022-02-26 To retain details while removing artifacts caused by dxt compression on textures Sample
1x_artifacts_bc1_free_alsa.pth Alsa 1x CC BY-NC-SA 4.0 ESRGAN BC1 Compression BC1 take 2
1x_BCGone_Smooth_110000_G Mutin_Choler 1x WTFPL ESRGAN BC1 Compression 2020-08-02 Attempts to remove the damages done by BC1 compression. 1xESRGAN https://imgsli.com/MjU5MzQ/4/3
1x_BCGone-DetailedV2_40-60_115000_G Mutin_Choler 1x WTFPL ESRGAN BC1 Compression 2021-03-18 1xESRGAN
Dithering
Model Name Author Scale License Architecture Purpose (short) Date Posted Purpose (Full) Pretrained_Model_G Sample


1x_artifacts_dithering_alsa.pth (pcloud)(mega) Alsa 1x CC BY-NC-SA 4.0 ESRGAN Dithering Dithered Images JPG (0-20%)
4xFSDedither Jacob 4x GNU GPL3 ESRGAN Dithering 2020-01-29 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 (https://github.com/ManuelFritsche/real-world-sr/tree/master/esrgan-fs/codes) for better details compared to plain ESRGAN. RRDB_ESRGAN_x4 https://imgsli.com/MTI1NTY/
4xFSDedither_Manga Jacob 4x GNU GPLv3 ESRGAN-FS Dithering 2020-04-21 Cartoons/pixel art/other non-realistic stuff with dithering RRDB_ESRGAN_x4.pth https://imgsli.com/MTQ3Nzc
4xFSDedither_Riven Jacob 4x GNU GPLv3 ESRGAN-FS Dithering 2020-07-11 Fine-tuned 4xFSDedither 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. 4xESRGAN https://imgsli.com/MTg5NTM

https://imgsli.com/MTg5NTQ

https://imgsli.com/MTg5NTU

https://i.imgur.com/j7Wtn0G.png

1x_DitherDeleter-Smooth_104000_G Mutin_Choler 1x WTFPL ESRGAN Dithering 2021-01-23 1xESRGAN https://imgsli.com/Mzg0MTk/1/2
1x_DitherDeleterV3-Smooth Mutin_Choler 1x WTFPL ESRGAN Dithering 2021-04-18 Attempts to remove the damages done by dithering. For this model, I downscaled all of the HR images to make their pixel size closer to 1000x1000 using the Box filter and then downscaled them again by 50% using the Point filter. Afterwards, I applied 32-bit Riemersma to every image in the dataset. 1xESRGAN https://imgsli.com/NTA5MjU/1/0

Blurring

Model Name Author Scale License Architecture Purpose (short) Date Posted Purpose (Full) Pretrained_Model_G Sample
1x_ThePi7on-Solidd_Deborutify_UltraLite_260k_G ThePi7on 1x CC BY-NC-SA 4.0 ESRGAN Deblur 2021-05-09 Sharpening, line darkening and slight line thinning, specifically made for the Boruto anime.
1x_DeBLR BlackScout 1x Unknown or None ESRGAN Deblur 2020-02-19 General Deblurring
mdeblur LyonHrt 1x UNKNOWN ESRGAN Blurring Strong deblurring model
1x_PixelSharpen_100000 DinJerr 1x CC BY-NC-SA 4.0 ESRGAN Pixel Art 2021-03-19 Restores blurry/upscaled pixel art. 1xESRGAN https://imgsli.com/NDUxNDQ/5/4
1x_ArtClarity DinJerr 1x WTFPL ESRGAN Pixel Art 2021-08-05 Texture retaining denoiser and sharpener for digital artwork. Helps resized/generated artwork look more like it is in 'native' resolution. 4xPSNR Sample Gallery
1x_Fatality_DeBlur Twittman 1x CC BY-NC-SA 4.0 ESRGAN Deblur 2019-10-16 Dataset: Mix of anime, manga, and photos 1x_DeJpeg_Fatality_01_175000_G.pth
UnResize V3 (1x_UnResize_V3_200000_G.pth) Twittman 1x CC BY-NC-SA 4.0 (?) ESRGAN Deblur / Unresize Dataset: Anime, manga, and photos -

Purpose: Fix images that have been arbitrarily / poorly resized, such as non-integer nearest-neighbor upscaling/downscaling - Also acts as an image sharpener/deblur when used on slightly soft inputs

1x_UnResize_MKII_030000_G.pth
1x-Focus and 1x-Focus_Moderate Kim2091 1x CC BY-NC-SA 4.0 ESRGAN Deblur These models deblur most images. It was trained mostly on aniso2 and iso blurring (BSRGAN augmentation) with some gaussian mixed in. It performs well on most blurry images, but I'd recommend using something like Fatality_Deblur for very strong gaussian blur. Dataset: The UniScale Dataset + My Fabric dataset 1xESRGAN Sample 1

Sample 2

Sample 3

Sample 4

ReFocus V3 (1x_ReFocus_V3_140000_G.pth) Twittman 1x CC BY-NC-SA 4.0 ESRGAN Deblur / ReFocus 2022-03-12 Dataset: Photos, Anime, and manga -

Purpose: DeBlur, ReFocus, Sharpen Real life style images, but will work on Anime images too.

1x_Saiyajin_DeJPEG_300000_G.pth
ReFocus Cleanly (1x_ReFocus_Cleanly_100000_G.pth) Twittman 1x CC BY-NC-SA 4.0 ESRGAN Deblur / ReFocus 2022-01-21 Dataset: Anime, manga, and photos -

Purpose: DeBlur, ReFocus, Sharpen Manga, Anime and cartoon style images, but will work on real life images too.

1x_ReFocus_V3_110000_G.pth

Banding

Model Name Author Scale License Architecture Purpose (short) Date Posted Purpose (Full) Pretrained_Model_G Sample
1x_N64clean BlueAmulet 1x CC BY-NC 4.0 ESRGAN Banding 2020-07-16 N64 textures use a color depth of 5-bits per channel, this model attempts to clean them, restoring smooth gradients in textures 1x_BC1-smooth2.pth
1x_Bandage-Smooth Mutin_Choler 1x WTFPL ESRGAN Banding 2021-04-16 Attempts to remove the damages done by color banding. For this model, I downscaled all of the HR images to make their pixel size closer to 1000x1000 using the Box filter and then downscaled them again by 50% using the Point filter. Afterwards, I applied 64-bit color banding to every image in the dataset. 1xESRGAN https://imgsli.com/NTA1NTk/1/0
Debandurh_FS Ultra-lite (1x_Debandurh_FS_lite_140000_G.pth) Twittman 1x CC BY-NC-SA 4.0 (?) ESRGAN Debanding images Dataset: Anime, real life, manga None

Halo Removal

Model Name Author Scale License Architecture Purpose (short) Date Posted Purpose (Full) Pretrained_Model_G Sample
1x_DeEdge_105000_G Mutin_Choler 1x WTFPL ESRGAN Halo remover 2020-11-03 1xESRGAN https://imgsli.com/MjgxMDU/3/0

Noise

Model Name Author Scale License Architecture Purpose (short) Date Posted Purpose (Full) Pretrained_Model_G Sample
LADDIER1 Alexander Syring 4x Unknown ESRGAN Denoise 2019-11-13 Remove noise, grain, box blur, lens blur and gaussian blur and increase overall image quality. RRDB_ESRGAN_x4.pth
1x_NoiseToner-Poisson-Detailed_108000_G Mutin_Choler 1x WTFPL ESRGAN Noise remover 2021-10-08 Attempts to remove the damages done from noise. Successor of sorts to Noisetoner_Poisson_150000_G 1xESRGAN https://imgsli.com/NzQ5MzM/2/0
1x_NoiseToner_Poisson_150000_G Mutin_Choler 1x WTFPL ESRGAN Noise remover 2020-10-14 1xESRGAN https://imgsli.com/MjU1ODU/3/1
1x_NoiseToner_Uniform_100000_G Mutin_Choler 1x WTFPL ESRGAN Noise remover 2020-10-14 1xESRGAN https://imgsli.com/MjU1Nzg/2/1
1x_ISO_denoise_v1 Alpha 1x WTFPL ESRGAN denoise not on discord Remove high ISO noise 1xESRGAN
1x_ISO_denoise_v2 Alpha 1x WTFPL ESRGAN denoise not on discord Remove high ISO noise ISO denoise v1
Film-Degrainer_1-000 Tika 1x CC0 ESRGAN denoise not on discord Remove film grain/noise none
1x_Fatality_NoiseToner DinJerr 1x WTFPL ESRGAN sharpen & denoise 2023-04-16 Interpolation of Mutin_Choler's various 1x_NoiseToner_Poisson with Twittman's 1x_Fatality_DeBlur none

Oversharpening

Model Name Author Scale License Architecture Purpose (short) Date Posted Purpose (Full) Pretrained_Model_G Sample
1x_DeSharpen Loinne 1x CC BY-NC-SA 4.0 ESRGAN Denoise 2019-06-03 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 oversharpening, 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 oversharpened, for example in image consisting of several combined images, 1 of them with sharpen noise. 1st attempt on random sharpening with the same dataset at 200000 iterations, which was trained on non-random desharp model, total ~600000 iterations on 3 models.

DeToon

Model Name Author Scale License Architecture Purpose (short) Date Posted Purpose (Full) Pretrained_Model_G Sample


detoon LyonHrt ESRGAN Detooning 2019-06-24 A toon to realistic shading style model to wiki under drawings
1xDoubleDetoon Joey 1x CC BY-NC-SA ESRGAN Detooning 2020-07-31 An attempt to detoon images/drawings of people 1xESRGAN



Image De/Colorization

Model Name Author Scale License Architecture Purpose (short) Date Posted Purpose (Full) Pretrained_Model_G Sample
BS_Colorizer/Vapourizer BlackScout 1x GNU GPLv3 ESRGAN Colorization 2020-01-30 B/W | 100% Desaturated images. It mostly results in Blue and Yellow images with slight hints of Green, Orange and Magenta. You are free to use this as a pretrain to achieve better results. 1xESRGAN
1x_SpongeColor-Lite Joey 1x CC BY-NC-SA 4.0 ESRGAN (lite) [nf=32 nb=12] Colorization 2020-10-15 The first attempt at ESRGAN colorization that produces more than 2 colors. Doesn't work that great but it was a neat experiment. none
Deoldify Rastrum 4x CC0 ESRGAN Photos 2019-08-29 Old black and white photo restoration. Falcon Fanart

Stylization

Images

Model Name Author Scale License Architecture Purpose (short) Date Posted Purpose (Full) Pretrained_Model_G Sample
Rek's Effeks Photoanime v2 Rek 4x GNU GPL3 ESRGAN Stylization 2020-10-17 Photo stylization from JPEGs. Trained on images upscaled by ISO Denoise v2 -> DeJPEG Fatality PlusULTRA -> NMKD Yan2. Essentially a combination of that model chain into one. Use if you're looking for a stylized output, not photo quality. NMKD Yandere2
Ghibli_Grain.pth nonogamester#3975 1x CC BY-NC-SA 4.0 ESRGAN Realistic Ghibli Grain 2022-11-16 Text or Anime. Attempt to get the nostalgic grain feel of classically animated Ghibli movies. Got the idea from researching the making of Ronin's :PKT_blue: 1x_AnimeFilmGrain28k. This was made running Nate video denoising with the preset "more denoising" on kiki's delivery service then overlaying it 66.65 percent over the original with DaVinci. On digital drawn anime it gives a slightly [10%] more organic/sharp feel to black lines. Matches well with content that already has a light digital grain. Run Twice for heavy grain. Eternal Thanks to all that indulge all my :psyduck: Wonderings :element_fire_neon: https://slow.pics/c/M6dNTu6G

https://imgsli.com/MTM0Nzcw/9/8

1x_ReDetail_v2_126000_G DinJerr 1x WTFPL ESRGAN Digital Illustration 2021-02-11 A failed model that's supposed to insert details into paintings. It's actual use is to interpolate with other 1x models such as with 1x_DoubleDetoon to reduce its colour warping, or with 1x_ArtClarity to emphasize more on feature extraction. 1x_ArtClarity

Specialized Models

Text

Model Name Author Scale License Architecture Purpose (short) Date Posted Purpose (Full) Pretrained_Model_G Sample
2x_BSTexty BlackScout 2x GNU GPLv3 ESRGAN Text 2020-03-04 As the name might suggest, this model aims to upscale text with less distortion that other models. It seems to do a good job generally, but don't expect it to be a state of the art model that can upscale magazines and stuff. It makes things more readable but since it was train on B/W pictures it desaturates them. 2x_ESRGAN
4x-TextSharpV1 Kim2091 4x CC BY-NC-SA 4.0 ESRGAN Text 2021-12-28 Text or Anime Interpolation between 4x-UltraSharp and 4x-TextSharp-v0.5. This is a duplicate listing of 4x-AnimeSharp Sample 1

Sample 2


NMKD Typescale Nmkd 8x WTFPL ESRGAN Text 2020-11-04 Low-resolution text/typography and symbols 8xESRGAN

Inpainting

Model Name Author Scale License Architecture Purpose (short) Date Posted Purpose (Full) Pretrained_Model_G Sample
1x_sudo_inpaint_PartialConv2D_424000_G.pth sudo rm -rf / --no-preserve-root#8353 1x CC BY-NC-SA 4.0 ESRGAN Inpainting 2020-12-05 Experimental PartialConv2D attempt to paint with ESRGAN. Took ~10.4 days of training on a P100 and around 1.5 months in total due to Colab limits. Not sure if I will continue training it since training is very slow, but may get better.. Warning: Result can vary with different tilesizes. Try not to tile your data. None Samples
1x_NMKD-YandereInpaint_375000_G.pth Nmkd 1x Unknown ESRGAN Inpainting Unknown None Samples

Fabric/Cloth

Model Name Author Scale License Architecture Purpose (short) Date Posted Purpose (Full) Pretrained_Model_G Sample
4x-Fabric and Fabric-Alt Kim2091 4x CC BY-NC-SA 4.0 ESRGAN Fabric 2021-09-15 This model set upscales fabric or cloth textures (works on cats too!). The Alt model is just an earlier iteration version. It may work better on some images.The images need to be minimally compressed or passed through a decompression model first. It works with DDS compression though. 4xESRGAN Sample 1

Sample 2

Sample 3

Sample 4


Alphas

Model Name Author Scale License Architecture Purpose (short) Date Posted Purpose (Full) Pretrained_Model_G Sample
FireAlpha BlueAmulet 4x CC BY-NC 4.0 ESRGAN Alpha (4 channel) 2019-11-12 none
4x_1ch-Alpha-Lite Joey 4x CC BY-NC-SA 4.0 ESRGAN Alpha 2020-12-15 Obsoleted by Joey's Fork - Alpha channels of PNGs none
2x_Gen5-Alpha Joey 2x CC BY-NC-SA 4.0 ESRGAN Pixel Art with Tranparency / Alpha Channel 2021-02-03 4xFireAlpha
8x_Sphax-Alpha-NN Joey 8x CC BY-NC-SA 4.0 ESRGAN Pixel Art with Tranparency / Alpha Channel 2021-02-04 Replica of sphax with transparency 4xFireAlpha
4x_PocketMonsters-Alpha Joey 4x CC BY-NC-SA 4.0 ESRGAN Pixel Art with Tranparency / Alpha Channel 2021-02-04 upscaling pixel art with alpha channels that should perform better than any other. It should work well on both cartoon and 3D styled content. 4xFireAlpha
Skyrim Alpha Deorder 4x CC0 ESRGAN Pixel Art with Tranparency / Alpha Channel Dataset: Alpha Channels from Skyrim Unknown

CGI

Model Name Author Scale License Architecture Purpose (short) Date Posted Purpose (Full) Pretrained_Model_G Sample
1xRedImage10000.pth 3majsie1995 1x CC BY-NC 4.0 ESRGAN Correct old color photos that are tinted red 2022-08-04 Correct old color photos that are tinted red. Model only tested in nature. 1xESRGAN https://i.imgur.com/5O74rRL.png
WaifuGAN_v3_30000 DinJerr 4x CC BY-NC 4.0 ESRGAN CGI 2019-07-12 Upscaling CG-painted anime with variable outlines. Manga109v2.pth
2xBS_Wolly BlackScout 2x GNU GPLv3 ESRGAN CGI Animation 2020-02-07 Pixar Movies or Wall-E pictures/frames 2xESRGAN
4xSGI ChrisNonyminus 4x GNU GPL3 ESRGAN CGI 2020-12-17 Upscaling and dedithering pre-rendered sprites, images and textures made in the 90s. Basically vintage CGI. 4xESRGAN https://imgur.com/a/iwLTjeB
CGIMaster_v1 Madiator 2x GNU GPL3 ESRGAN CGI Animation 2021-01-14 mixed 3D/2D CGI animations on some 2D animations it might sharpen the edges. General usage is to upscale CGI animations compressed by YouTube. Sharp_Anime_v2 https://imgsli.com/Mzc1NTM

Luminance/Chroma

Model Name Author Scale License Architecture Purpose (short) Date Posted Purpose (Full) Pretrained_Model_G Sample


1x_BSLuma BlackScout 1x GNU GPLv3 ESRGAN Luma 2020-02-28 Fix Luminance issues?? LumaSharpen ESRGAN Edition? - This model mostly does what "Lumasharpen" algorithms do. It may help fixing images with Luminance images issues? Like old DVD rips? I am not sure, didn't test. none
1x_BSChroma BlackScout 1x GNU GPLv3 ESRGAN Sharpen 2020-02-29 ChromaSharpen - makes the colors slightly more vibrant with a sideeffect of possibly adding Chromatic Abberation to the image. I am not sure about the usage of this model on real case scenarios but anything blurry or with fuzzy colors could work none
1x_3mChroma 3majsie1995 1x CC BY-NC 4.0 ESRGAN Chroma 2023-01-04 Improves chromatic aberration. It's not as good as 1x_BSchroma, but I recommend you to test it! 1xESRGAN https://i.imgur.com/oSub859.png

Cats

Model Name Author Scale License Architecture Purpose (short) Date Posted Purpose (Full) Pretrained_Model_G Sample
Cat_Patch Twittman 4x CC BY-NC-SA 4.0 (?) ESRGAN Cats not on discord previous attempt

Coins

Model Name Author Scale License Architecture Purpose (short) Date Posted Purpose (Full) Pretrained_Model_G Sample
4x_Nickelfront BlackScout 4x GNU GPLv3 ESRGAN Coins 2020-02-13 Upscale coins. That's it. If you were mad at me because Nickelback doesn't make any sense. Now you have the perfect solution to your problems. If you want to upscale nickels or anything with similar texture made out of metal, now you can. It works pretty well for a joke. 4xESRGAN

Faces

Model Name Author Scale License Architecture Purpose (short) Date Posted Purpose (Full) Pretrained_Model_G Sample
4x_Faces_N_250000.pth Twittman 4x ESRGAN Faces 2019-07-18 Upscales images of faces of different scales, sometimes resulting in monsters. 4xESRGAN
4x_Faces_04_N_180000_G Twittman 4x ESRGAN Faces 2019-07-23 Upscale faces both pixelized and real 4x_Faces_N_250000.pth
Face-Ality V1 (4x_Fatality_Faces_310000_G.pth) Twittman 4x CC BY-NC-SA 4.0 (?) ESRGAN Faces Dataset: Custom (Faces) 4x_Faces_04_N_180000_G.pth
4x_SmolFace_200k DinJerr 4x CC BY-NC-SA 4.0 ESRGAN Art/People 2022-08-19 A sharp upscaler trained specifically for small sprite faces. Does not blend, so avoid using on painted/photo portraits unless you were trying to retain more of the outlines somehow. The _clean version has denoising/dedithering training on top of it. 4x_NMKD-UltraYandere_300k SmolFace: https://imgsli.com/MTIxNjU1/8/9

SmolFace_clean: Sample 1 Sample 2 Comparison with anime models

BigFace_v3, BigFace_V3_Blend, BigFace_V3_Clear DinJerr 4x CC BY-NC 4.0 ESRGAN Art/People 2020-07-22 Pixel art upscaler for faces drawn in digital painting style. Best for game portraits with multiple shades (not cel-shaded). 4x_BigFace
TGHQFace8x tg 8x GNU GPL3 ESRGAN Faces 2019-11-27 Upscales blurry 128px faces, usefull for enhancing that someone in a small picture. 8xESRGAN https://imgsli.com/OTM1NQ
FArtFace DinJerr 4x CC BY-NC 4.0 ESRGAN Faces 2019-11-28 Rather old pixel art face upscaler. I suggest using BigFace_v3 or SmolFace instead. Sometimes produces better results though.
4x_BS_SbeveHarvey BlackScout 4x GNU GPLv3 ESRGAN Faces 2020-07-24 Upscale Steve Harvey, but maybe other things, somehow?? 4xESRGAN
Face Focus LyonHrt 4x CC BY-NC-SA 4.0 ESRGAN Faces not on discord Face De-blur - slightly out of focus / blurred images of faces. It is aimed at faces / hair 4xPSNR
TGHQFace8x Torrentguy 8x [https://www.gnu.org/licenses/gpl-3.0.en.html GNU GPLv3] ESRGAN Face Upscaling 8xESRGAN

Skin

Model Name Author Scale License Architecture Purpose (short) Date Posted Purpose (Full) Pretrained_Model_G Sample


x1_ITF_SkinDiffDetail_Lite_v1 intheflesh#3116 1x CC BY-NC-SA 4.0 ESRGAN Skin Upscaling Adds plausible high frequency detail and removes subtle blur. This is an early unfinished attempt at a x1 Lite model designed specifically for enhancing detail on skin diffuse textures of 3d characters. Even in its current state it works quite well.

Best suited for uncompressed or cleaned textures - otherwise it may just enhance any existing compression artefacts too. The training set included faces, body parts, eyes and hair in a variety of skin types and tones so it should work well on most related diffuse textures.

However it's not just limited to skin, many other images and textures can be enhanced with this model. The results are subtle, so run multiple times if desired.

50/50 Interpolation of DIV2K-Lite and SpongeBC1-Lite


Foliage/Ground

Model Name Author Scale License Architecture Purpose (short) Date Posted Purpose (Full) Pretrained_Model_G Sample
1x_Plants_400000_G.pth Muf 1x Public Domain ESRGAN bad upscale 2021-06-23 Images of plants, trees or other foliage upscaled with Photoshop Preserve Details 2.0. Sharpens and "subdivides" details and noise so it doesn't look upscaled. 1x_NMKD-h264Texturize_500k.pth https://imgsli.com/NTg2MDk https://imgsli.com/NTg2MDg https://imgsli.com/NTg2MTA
Ground tldr_coder 4x WTFPL ESRGAN Upscales ground textures 2021-06-23 Dataset: Custom (Ground textures Google)

Video Games

Game Screenshots

Model Name Author Scale License Architecture Purpose (short) Date Posted Purpose (Full) Pretrained_Model_G Sample
Minepack BlackScout 4x CC BY-NC-SA ESRGAN Upscaling pack.png 2020-01-26 Upscales Minecraft screenshots by 4x. May suffer from haloing, weird patterns on blocks and JPEG-like artifacts. 4xESRGAN.pth
4x_Link Fielran#1024 4x GNU GPL3 ESRGAN Chainmail game textures. Alternatively, it can be used to turn plain images into chainmail. 2022-10-06 A highly generative model for chainmail game textures. Works fairly well on items without visible chainmail texture (ie just a grey area) or items with clear chainmail texture, less well on items with present but poorly defined chainmail texture. I'm not 100% happy with it but it is still an improvement over existing models in enough situations to be worth releasing. 4x_Nickelfront_14000G.pth Sample 1

Sample 2

Sample 3

4x-FatePlus-lite Kim2091 4x CC BY-NC-SA 4.0 ESRGAN-lite Anime PSP games, Fate Extra 2022-07-03 This model was trained as a favor to Demon and the Fate Extra community. It leaves a nice grain on the images and upscales lines and details accurately without looking odd. This model works on most anime-style PSP games. Enjoy! It works best on content with dithering and quantization. NOTICE: I have included both NCNN and ONNX models to make upscaling easier if you rely on either of these. For NCNN, there's two versions. One is FP16 and the other is FP32. FP16 works best on RTX GPUs. Choose FP32 if in doubt about compatibility, or if FP16 doesn't work for you. To use ONNX, download chaiNNer and upscale through there with the ONNX nodes. 4x-AnimeSharp-lite
4x_GameAI_2.0 4x_GameAI_1.0 Tal 4x WTFPL ESRGAN PS2 textures, cartoonish and realistic game textures. 2022-02-27 This model is intended to mainly handle PS2 compression and a mixture of Realistic and cartoonish textures, it's not meant to be used for very low resolution textures such as item icons. Dataset: I used textures from, "(A Hat in Time, Kingdom Hearts 3 and World of Final Fantasy) for GameAI_2.0" "(Skyrim, The Witcher 3, Resident Evil 4, Final Fantasy XV and Tales of Vesperia ) for GameAI_1.0" Dataset_size: 21.9k to 70k textures , up to 1024x1024 per texture 4x_GameAI_1.0 Sample 1

Sample 2

Sample 3

Sample 4

4xPackCraft_v4 Joey 4x CC BY-NC-SA 4.0 ESRGAN Upscaling pack.png 2020-09-05 Designed to upscale one specific minecraft screenshot. Results on dissimilar screenshots may be poor. 4xESRGAN/previous versions Sample
MinecraftAlphaUpscaler with Good data Washed Up 4x GNU GPLv3 ESRGAN Upscaling pack.png 2020-01-29 Designed to upscale one specific minecraft screenshot. Results on dissimilar screenshots may be poor. RRDB_PNSR_x4.pth
ScreenBooster V2 BlackScout 4x CC BY-NC-SA ESRGAN CGI 2020-01-29 Game Screenshots none
BS_ScreenBooster_SPSR BlackScout 4x CC BY-NC-SA SPSR CGI 2020-07-18 This model is designed to upscale game screenshots (3D Games) by 4 times. The SPSR version is an improvement over the ESRGAN based V2. Screenbooster V2

Normal Map/Bump Map Generation

You may want to use this instead: https://github.com/JoeyBallentine/Material-Map-Generator

Model Name Author Scale License Architecture Purpose (short) Date Posted Purpose (Full) Pretrained_Model_G Sample
1x_NormalMapGenerator-CX-Lite Joey 1x CC BY-NC-SA 4.0 ESRGAN Lite [nf=32 nb=12] Map Generation - normal maps 2020-11-04 Generating normal maps from textures 1x_DIV2K-Lite_450k.pth
1x_FrankenMapGenerator-CX-Lite Joey 1x CC BY-NC-SA 4.0 ESRGAN Lite [nf=32 nb=12] Map Generation - roughness and displacement maps 2020-11-07 This model generates "Franken Maps" (named after Frankenstein), which is a custom material map combination I made. Basically, the Red channel of RGB is just the texture converted to grayscale, the Green channel is the roughness map, and the Blue channel is the displacement map. I had to do this to get around the current limitation of CX loss where it requires a 3 channel output (otherwise I would have just made a 2 channel model, or separate single channel models). As of right now the channels need to be manually split from each other but I will be making a tool for doing this automatically in the coming days. 1x_DIV2K-Lite_450k.pth
1x_normals_generator_general_215k LyonHrt 1x CC BY-NC-SA 4.0 ESRGAN Map Generation - Normal Maps This model generates "Franken Maps" (named after Frankenstein), which is a custom material map combination I made. Basically, the Red channel of RGB is just the texture converted to grayscale, the Green channel is the roughness map, and the Blue channel is the displacement map. I had to do this to get around the current limitation of CX loss where it requires a 3 channel output (otherwise I would have just made a 2 channel model, or separate single channel models). As of right now the channels need to be manually split from each other but I will be making a tool for doing this automatically in the coming days. none (no interpolation)

Video Game Textures

Model Name Author Scale License Architecture Purpose (short) Date Posted Purpose (Full) Pretrained_Model_G Sample


Fallout Weapons (Fallout 4 Weapons?) Bob 4x CC BY-NC-SA 4.0 ESRGAN Game textures 2019-08-05 Video game textures, mostly metal rusty, clean or painted Manga109Attempt
Fallout Weapons V2 Bob 4x CC BY-NC-SA 4.0 ESRGAN Game textures 2019-08-23 Video game textures, mostly metal rusty, clean or painted Fallout 4 Weapons
4x_ThiefGoldMod_100000 Akven 4x no idea ESRGAN Game textures 2020-05-15 Version of the previous model but based on Manga109 pretrained model and with slightly different dataset. Sometimes gives better results especially for wood and metal, sometimes worse. Sometime generates the same dotted artifacts on very bright/white images. 4x_Manga109Attempt https://imgsli.com/MTYzNjI
4x_ThiefGold_110000 Akven 4x no idea ESRGAN Game textures 2020-05-15 Various game textures. Primary wood, metal, stone RRDB_ESRGAN_x4 https://imgsli.com/MTYzNjI
4x-VolArt Kim2091 4x CC BY-NC-SA 4.0 ESRGAN Game textures/Art 2021-09-18 This model upscales artwork for the game Volfoss (2001). The model peaked in quality very quickly. The NR model removes most noise, but has the downside of removing transparent portions. Use the main model in most cases. 4xESRGAN Sample 1 Sample 2
2xFaithfulSPSR Joey 2x CC BY-NC-SA SPSR Pixel art 2020-07-19 Mainly just a test for SPSR. Seems to work better than the original 2xFaithful32_1316 that I used as a pretrained, even though it uses the same dataset. 2xFaithful32_1316
2x_FakeFaith-Lite Joey 2x CC BY-NC-SA 4.0 ESRGAN (lite) Pixel Art 2020-10-12 An attempt at recreating the "faithful" style without using the faithful dataset -- aka keeping the "pixel art" style of pixel art. 2x_Faithful-Lite
2x_Faithful-Lite Joey 2x CC BY-NC-SA 4.0 ESRGAN (lite) Pixel Art 2020-10-12 A "lite" model version of my Faithful model none
4x_HDCube3 Venomalia 4x CC BY-NC-SA 4.0 ESRGAN Gamecube and Wii textures. 2022-12-13 It can be used for all image formats supported by Gamecube and Wii hardware and can remove its typical artifacts like CMPR Block Compression (DXT1 algorithm, also known as BC1), color palette errors, color reduction up to 8bit color depth and 1bit alpha depth. 4x_HDcube2 https://imgsli.com/MTM5ODE0/0/1

Sample Gallery

4x_HDCube Venomalia 4x CC0-1.0 ESRGAN Gamecube and Wii textures. 2022-05-10 Gamecube and Wii textures (mainly DXT and 8bit color compression). Is good for preserving fine details without affecting the original style too much, it is not suitable for pixel art, small icons and text under 16 pixel. 4x_NMKD Siax https://imgsli.com/MTA3NDQ3

https://slow.pics/c/rmQsQ6yb

1x_DXTless_SourceEngine_170000_G Xeller 1x WTFPL ESRGAN DTX5 Compression 2021-04-10 This model is made for Source Engine textures. It tries to remove compression artifacts such as blockyness, discoloration, green tint. It does pretty well on a lot of things, realistic stuff as well, but it was mostly made to work on TF2 textures. It's made to keep as much detail as possible, without any unnecessary denoising/sharpening. Huge thanks to Twittman for assisting me along this journey. 1x_Saiyajin_DeJpeg_300000_G https://imgsli.com/NDk1NTk

https://imgsli.com/NDk1NjA

https://imgsli.com/NDk1NjE

https://imgsli.com/NDk1NjI

Trixie LyonHrt 4x CC BY-NC-SA 4.0 ESRGAN Faces / Game Textures not on discord 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 none
Skyrim Armory Alsa 4x CC BY-NC-SA 4.0 ESRGAN Game textures/equipment not on discord Manga109Attempt
Skyrim Misc Deorder 4x CC0 ESRGAN Game textures not on discord Skyrim Diffuse Textures ?
Skyrim Wood Laeris 4x ESRGAN Game textures not on discord Wood ?
4x-SkyrimTexV2.1 and V2_Fabric Kim2091 4x CC BY-NC-SA 4.0 ESRGAN Game textures 2021-09-25 This model set upscales most Skyrim textures. I hope it helps 🙂 The base model (2.1) works well on stone, wood, metals, and most other textures. The Fabric model is a 50/50 interpolation with a stronger iteration of this model and my Fabric model. As you can tell by the name, it's intended for Fabric textures. If you want to upscale Alpha textures, use a model dedicated to it. There's an INFO file in the folder, it just explains the extra models 4xESRGAN Sample 1

Sample 2

Sample 3

Sample 4

Forest LyonHrt 4x CC BY-NC-SA 4.0 ESRGAN Game textures not on discord Wood / Leaves none
Morrowind Mixed mkultra 4x CC BY-NC-SA 4.0 ESRGAN Game textures not on discord Morrowind Mod Textures 4xESRGAN
Morrowind 2.0 mkultra 4x CC BY-NC-SA 4.0 ESRGAN Game textures not on discord Morrowind Mod Textures 4xESRGAN
Map LyonHrt 4x CC BY-NC-SA 4.0 ESRGAN Game textures not on discord Map / Old Paper with text none

Video Restoration

Video Compression

Model Name Author Scale License Architecture Purpose (short) Date Posted Purpose (Full) Pretrained_Model_G Sample
1x_cinepak_200000 Twittman 1x CC BY-NC-SA 4.0 (?) ESRGAN Compression 2019-07-03 Removal of Compression such as Cinepak, msvideo1 and Roq none
1x_cinepak_alt.pth buildist 1x GNU GPLv3 ESRGAN Compression Cinepak (only) none
DeBink v4/v5/v6 Kim2091 1x CC BY-NC-SA 4.0 ESRGAN Compression 8.6.21 This model removes early 2000s Bink and other compression artifacts. Works well on almost any image or video compression type. none
DeBink_Lite Kim2091 1x CC BY-NC-SA 4.0 ESRGAN-lite Compression 2021-10-30 This model was trained on lossless video frames of Metal Arms: Glitch in the System, compressed with Bink V1 for the LR frames. It's a lot more efficient than my first DeBink model, and also has less artifacts. It's not quite as robust, but the compression is barely noticeable in videos after processing with this. none Video Sample
1x-RoQ_nRoll Kim2091 1x CC BY-NC-SA 4.0 ESRGAN Compression 2021-11-21 This model decompresses images and video compressed using RoQ. Config and presets will be added when Mega decides to let me use their site! none Sample 1

Sample 2

Sample 3

Sample 4

Sample 5

1x_DeRoqBeta-lite Kim2091 1x CC BY-NC-SA 4.0 ESRGAN Compression 2022-07-26 Incomplete lite model to remove ROQ compression Sample
NMKD h264Texturize Nmkd 1x WTFPL ESRGAN Texturizing 2020-10-20 Tries to reverse heavy h264 compression. Fails. Can be used to texturize images though. 4x ESRGAN
1x_Filmify4K_v2_325000_G.pth Muf 1x Public Domain ESRGAN artifact 2021-07-19 This model attempts to make films upscaled to 4K with Topaz Gaia-HQ look more natural and filmic. It sharpens, adds film grain, and smooths out small artefacts from the upscaling process. I recommend adding a tiny amount of grain to the input to seed the model (you can do this in VEAI), otherwise the film grain will remain static across frames that don't move much. Pretrain model used with permission to relicense from Twittman. 1x_UnResize_V3_110000_G.pth https://imgsli.com/NjE5MTE

https://imgsli.com/NjE5MTI

https://imgsli.com/NjE5MTM

https://imgsli.com/NjE5MTQ

DeIndeo (mirror) DeIndeo Wild West Quest#7975 4x WTFPL ESRGAN Indeo Compression Artifacts Dataset: Custom 4xESRGAN

VHS Tapes

Model Name Author Scale License Architecture Purpose (short) Date Posted Purpose (Full) Pretrained_Model_G Sample
1xBaldrickVHSFix_180000_G_V0.2 NimRodZorg 1x CC BY-NC-SA 4.0 ESRGAN VHS 2021-03-03 Fixing minor VHS Chroma and Pattern Noise - NOTE: only works on deinterlaced sources 1xESRGAN
ToonVHS Redslam 1x CC BY-NC-SA 4.0 ESRGAN VHS 2022-02-06 Best when used on cartoons, it can work on anime. Due to the dataset it does struggle a bit with orange colors and grainy dark spots. This model is meant to be used to clean up the image before using it on a 2x or 4x model. 1xESRGAN Examples of ToonVHS + other models

Examples of ToonVHS by itself

2x_VHS-upscale-and-denoise_Film_477000_G Itewreed 2x CC BY-NC-SA 4.0 ESRGAN VHS 2021-03-28 VHS captures of Film material, but may work on VHS recorded native SD-TV material as well. Also useable for cleaned up source material none
VHS-Sharpen-1x RTX 2080 ti hoarder 1x CC BY-NC-SA 4.0 ESRGAN VHS 2021-03-19 Make old VHS footage crispy. This model will not work on video and images with noticeable JPEG/Video compression artifacts, noticeable interlacing or haloing, heavy tape distortion/artifacts and scenes with tons of detail. For best results, use a downscaled HD capture of the VHS tape you intend to use it on. 1xESRGAN https://imgsli.com/NDUxNDY/2/3

Model Collections

Model Name Author Scale License Architecture Purpose Date Posted Description Sample
OptimusPrimal's Collection OptimusPrimal Various WTFPL ESRGAN Collection 2022-02-28 I'm dumping all my models here. (I moved my favorites to a favorite folder). Most of these models work well on DVD resolution animation sources. These are all ESRGAN models. Everything is with the WTFPL license. So do whatever you want with them. Credit is not needed, but would be nice if you use them to make something. There's a few that are interpolations with other user's publicly available models, or interpolations of other user's publicly available models (so technically, these are not mine), so credit goes out to them, the only ones I can remember off the top of my head is @twittman with the Fatality model series, which are great, and @cd VSGAN && poetry install with the Sol Levante/American Dad2 models, and if you do 50/50 with those 2 models, you get a really good model that I called Soladad.
Kim's Collection Kim2091 Various CC BY-NC-SA 4.0 Various (ESRGAN and ESRGAN-lite) Collection 2022-02-28 My (Kim's) main folder for hosting models. Every model of mine on this database links to subsections of this folder. If a link is broken or you just want to see nearly all of the models I've trained, look here.
UltraMix Collection Kim2091 Various CC BY-NC-SA 4.0 Various (ESRGAN and ESRGAN-lite) Collection 2022-06-28 This is a mixture of models based around UltraSharp and my other available models. These are usually interpolations that have separate but very helpful uses. As an example, UltraMix_Restore is a combination of UltraSharp and UniScale_Restore, and is great for video game textures.
NMKD's Collection NMKD Various Various Various (ESRGAN and ESRGAN-lite?) Collection 2022-02-28 NMKD's main folder for hosting models. Nearly every one of NMKD's models posted on this wiki link to it.
Zarxrax's Collection Zarxrax Various Various Compact Collection 2023-01-20 All of my compact models, including a few variations which might not be linked from elsewhere.
Joey's ESRGAN Bot Collection Various Various Various Various Collection 2022-08-04 Here is a dump of all the ESRGAN-Bot models. As it was a un-curated collection, many of the models here were never officially released, but I do not know exactly which ones. If one of your models is included here and you would like me to remove it, please let me know.


Pretrained Models

Looking for official models? Look here: https://upscale.wiki/wiki/Official_Research_Models

Model Name Author Scale License Architecture Purpose (short) Date Posted Purpose (Full) Pretrained_Model_G Sample
1xESRGAN victorca25 1x Apache License 2.0 ESRGAN Pretrained 2019-07-06 RRDB_ESRGAN_x4.pth
2xESRGAN victorca25 2x Apache License 2.0 ESRGAN Pretrained 2019-07-06 RRDB_ESRGAN_x4.pth
4xESRGAN xinntao 4x Apache License 2.0 ESRGAN Pretrained 2019-07-06 RRDB_ESRGAN_x4.pth
8xESRGAN victorca25 8x Apache License 2.0 ESRGAN Pretrained 2019-07-06 RRDB_ESRGAN_x4.pth
16xESRGAN victorca25 16x Apache License 2.0 ESRGAN Pretrained 2019-07-06 RRDB_ESRGAN_x4.pth
1x PSNR Pretrained Model BlueAmulet 1x Apache License 2.0 (ESRGAN's license) ESRGAN Pretrained 2020-04-20 The original RRDB_PSNR_x4.pth model converted to 1x, 2x, 8x and 16x scales, intended to be used as pretrained models for new models at those scales. These are compatible with victor's 4xESRGAN.pth conversions RRDB_PSNR_x4.pth
2x PSNR Pretrained Model BlueAmulet 2x Apache License 2.0 (ESRGAN's license) ESRGAN Pretrained 2020-04-20 The original RRDB_PSNR_x4.pth model converted to 1x, 2x, 8x and 16x scales, intended to be used as pretrained models for new models at those scales. These are compatible with victor's 4xESRGAN.pth conversions RRDB_PSNR_x4.pth
4x PSNR Pretrained Model BlueAmulet 4x Apache License 2.0 (ESRGAN's license) ESRGAN Pretrained 2020-04-20 The original RRDB_PSNR_x4.pth model converted to 1x, 2x, 8x and 16x scales, intended to be used as pretrained models for new models at those scales. These are compatible with victor's 4xESRGAN.pth conversions RRDB_PSNR_x4.pth
8x PSNR Pretrained Model BlueAmulet 8x Apache License 2.0 (ESRGAN's license) ESRGAN Pretrained 2020-04-20 The original RRDB_PSNR_x4.pth model converted to 1x, 2x, 8x and 16x scales, intended to be used as pretrained models for new models at those scales. These are compatible with victor's 4xESRGAN.pth conversions RRDB_PSNR_x4.pth
16x PSNR Pretrained Model BlueAmulet 16x Apache License 2.0 (ESRGAN's license) ESRGAN Pretrained 2020-04-20 The original RRDB_PSNR_x4.pth model converted to 1x, 2x, 8x and 16x scales, intended to be used as pretrained models for new models at those scales. These are compatible with victor's 4xESRGAN.pth conversions RRDB_PSNR_x4.pth
Compact Pretrained Models Zarxrax 1x-4x WTFPL Real-ESRGAN "compact" Pretrained 2022-07-31 This is a collection of pretrained models for Real-ESRGAN's Compact architecture. There are 1x, 2x, and 4x models, as well as 1x and 2x "UltraCompact" and "SuperUltraCompact" models (think of these as the equivalent to ESRGAN "lite" models). By using these are pretrains for your models, you can ensure that your models are able to be interpolated with other Compact models that were trained from these. These pretrains are compatible with most existing compact models.

Pretrained Discriminators

Model Name Author Scale License Architecture Purpose (short) Date Posted Purpose (Full) Pretrained_Model_G Sample
1x Pretrained Discriminator Pack Joey 1x CC BY-NC-SA 4.0 VGG Pretrained discriminators 2020-10-26 Most of these are my spongebob dataset so they will be more useful for cartoons but I did include my original faithful model discriminator. none
2x Pretrained Discriminator Pack Joey 2x CC BY-NC-SA 4.0 VGG Pretrained discriminators 2020-10-26 Most of these are my spongebob dataset so they will be more useful for cartoons but I did include my original faithful model discriminator. none
4x Pretrained Discriminator Pack Joey 4x CC BY-NC-SA 4.0 VGG Pretrained discriminators 2020-10-26 Most of these are my spongebob dataset so they will be more useful for cartoons but I did include my original faithful model discriminator. none
8x Pretrained Discriminator Pack Joey 8x CC BY-NC-SA 4.0 VGG Pretrained discriminators 2020-10-26 Most of these are my spongebob dataset so they will be more useful for cartoons but I did include my original faithful model discriminator. none
4x_RRDB-G_ResNet-D (Both G and D) Joey 4x CC BY-NC-SA 4.0 ESRGAN (G) / ResNet (D) Pretrained 2021-01-17 Clean bicubic downscales / Pretrained models (G/D) RRDB_ESRGAN_x4_old_arch.pth


Official Models

To cut down on the length of this page and reduce confusion, these models have been moved here: https://upscale.wiki/wiki/Official_Research_Models This page contains models such as: ESRGAN, BSRGAN, Real-ESRGAN, and more.

SOFVSR Models

These models are for SOFVSR, they will not work in Cupscale or the main ESRGAN forks. You need this to use it: https://github.com/JoeyBallentine/Video-Inference

Model Name Author Scale License Architecture Purpose (short) Date Posted Purpose (Full) Pretrained_Model_G Sample
3x_Video_TSSM Joey 3x CC BY-NC-SA 4.0 SOFVSR Animation 2020-11-17 none
3x_Video_TSSM_RRDB Joey 3x CC BY-NC-SA 4.0 SOFVSR RRDB / VESRGAN Animation 2020-11-25 none
2x_SBS11-RRDB Joey 2x CC BY-NC-SA 4.0 SOFVSR RRDB / SOFVESRGAN Animation 2020-12-03 none
4x_REDSVAL-7f-RRDB-Lite (both G and D) Joey 4x CC BY-NC-SA 4.0 SOFVSR RRDB / SOFVESRGAN Lite [nf=32, nb=12] IRL videos 2020-12-03 none
SOFVSR_REDS_F3_V1 Sunseille 4x WTFPL RRDB IRL videos 2021-02-26 upscale IRL videos. Use not recommended by creator none Sample Video
2x_VimeoScale_Unet Sazoji 2x CC BY-SA 4.0 SOFVSR-RRDB (light arch, 3 frames, nf=32, nb=12) Upscaling video content 2021-09-09 This model is meant to surpass VEAI 2x while being efficient to run quickly with fp16. The real-esrgan/BSRGAN augmentation and Unet should help with videos where the resolution is not ideal and can reconstruct details without effecting blurs in most cases. This model SHOULD run faster than real-esrgan while matching the resolving and enabling some multiframe feature extraction. No major denoising/compression/blurring effects (or artifacts) should be found. Self-Trained Base for the Unet finalisation with vgg_fea discrim Sample 1 Sample 2 Sample 3
4x_VimeoScale Sazoji 4x CC BY-SA 4.0 SOFVSR-RRDB (5 frames) Upscaling IRL video content 2021-10-08 This model is one of the longest I have trained and tuned, includes some noise training but this is not intended for deblocking/decompression. Combined with https://github.com/JoeyBallentine/Video-Inference 's fp16 mode, this should handle most SD and 720p content at or faster than ESRGAN, with a significant bump in quality. Self-Trained Base for the Unet finalisation with vgg_fea discrim Sample 1 Sample 2 Sample 3 Sample 4

Pretrained SOFVSR Models

Model Name Author Scale License Architecture Purpose (short) Date Posted Purpose (Full) Pretrained_Model_G Sample
video_G victorca25 4x Unknown SOFVSR net Trained on REDS' size dataset none
VESRGAN_G victorca25 4x Unknown RRDB Trained on REDS' size dataset none

RIFE Models

Model Name Author Scale License Architecture Purpose Date Posted Description Pretrained_Model_G Sample
sudo_rife4_269.662_testV1_scale1.pth sudo 2x CC BY-NC-SA 4.0 RIFE Animation interpolation 2022-06-25 I never really mentioned it in model-releases prior since I think not too many care about interpolation here, but I trained a rife4 model for animation a some months ago, which is better than rife4 and rife4.2 imo. Thought I should also mention it here as well. I also converted it into ncnn. (Nihuis rife ncnn models are only exported with the fastest mode and not the best quality. I exported ncnn models for the most important quality settings. Due to different export/quality settings, there are multiple models. For that reason alone, my ncnn models are much better too, since nihui only exported the fastest one.) My https://github.com/styler00dollar/VSGAN-tensorrt-docker also has the rife ncnn extention, which can use VMAF, dedup, scene detection and so on, which I would recommend. My models are in that extention as well, just select model 10, 11 or 12 and use the dev docker. That test video is done with 2x framerate, enbemble True and FastMode False, combined with scene detection and dedup stuff, tta False. Towards the best quality rife can do. Plz don't steal without credits, k thx. Sample Video

CAIN Models

CAIN is a video frame rate interpolation AI. These models are intended for use with CAIN only. These CANNOT be used with Cupscale or ESRGAN, you need this: https://gitlab.com/hubert.sontowski2007/cainapp
Model Name Author Scale License Architecture Purpose Date Posted Description Pretrained_Model_G Sample
explodV1 hubert 2x BSD-3-Clause CAIN YUV Animethemes 2022-07-29 I think one of sharpest models for anime

Plz don't steal without credits, k thx.

Sample Video
190562_vimeo_enchanced.pth hubert 2x MIT CAIN RL videos - interpolation 2021-07-13
cainREALTIME hubert 2x MIT CAIN (CAIN 1 Group) RL videos - interpolation 2021-07-13 Trained mostly on abba music videos
cainliteanime.pth hubert 2x MIT CAIN Anime - interpolation 2021-08-06 Sample
rvpV1 (pcloud) rvpV1 (mediafire) sudo rm -rf / --no-preserve-root#8353 2x CC BY-NC-SA 4.0 CAIN Video Frame Interpolation for anime openings 2021-09-14 My ~~first~~ (ok technically second) attempt to create a video frame interpolation model and I like how it turned out. To use it, you can either use cain-App, Colab-CAIN (no longer available) or the bot in the Game Upscale discord (model called rvpv1 there, just use --model rvpv1). Some demo videos in pcloud, but you need to download them. The web player seems to playback in low fps. The architecture is mainly the same to the original CAIN, but i modified the padding to be zero padding instead. ``.pt`` means JIT model, `.pth`, means normal pytorch model. Architecture file is in pcloud as well. And no, no cupscale or flowframes. Dataset: Modified Animeinterp dataset Sample Video
cvpv6.pth hubert 2x MIT CAIN Anime - interpolation 2021-09-21

Waifu2x Models

Some helpful links for Waifu2x, rewritten in PyTorch. The repos contain the pretrained models themselves as well.