This is a complete (as of today) collection of hundreds of ESRGAN upscalers for all types of different applications. I have seen people uploading 1 upscaler at a time which is not only inefficient it spams the feed. So here is the answer, and its comprehensive!
There are a total of 327 upscalers running the gamut from CGI to anime to realism. Total uncompressed size is 20GB. They range from a variety of authors and from 1x to 16x upscale multipliers.
Some of you may know me from Reddit where I have posted about my data/model hoarding tendencies. Now I am putting it to good use! Enjoy!
Supplement 1:
Some additions to the original file that I found were omitted:
x1_ITF_SkinDiffDetail_Lite_v1.pth
TGHQFace8x_500k.pth
8xNMKDFaces160000G_v10.pt
8x NMKD-Superscale_150000_G.pth
4xUniscale-Strong [42400g].pth
4x-UniScale-Balanced [72000g].pth
4x-UltraSharp.pth
4x RealisticRescaler_100000_G.pth
1x Usul.v1_465000_G.pth
1x Plants_400000_G.pth
1x Filmify4K.v2_325000_G.pth
ORIGINAL VERSION EDIT: Here is is a list of all the upscaler filenames.
Todays Quick-Tip is open cmd window in the directory you want a list of files for then type: "dir /s/b/o:gn > filenames.txt"
1x_artifacts_bc1_free_alsa.pth
1x_artifacts_jpg_20_40_alsa.pth
1x_artifacts_jpg_40_60_alsa.pth
1x_Bandage-Smooth-[64]_105000_G.pth
1x_BCGone_V4.pth
1x_BroadcastToStudioLite_485k.pth
1x_BSChroma.pth
1x_BSDeBicubic.pth
1x_BSLuma.pth
1x_Debandurh_lite.pth
1x_DeBLR.pth
1x_DeBLRFS.pth
1x_DEDXT.pth
1x_DEHALO_LITE_V3_SOLIDD_296000_G.pth
1x_DeInterPaint_205000_G.pth
1x_DeteriorationFrames.pth
1x_DitherDeleter32_90000_G..pth
1x_DitherDeleter64_70000_G.pth
1x_DitherDeleter-Smooth_104000_G.pth
1x_DXTless_SourceEngine_170000_G.pth
1x_DynamicDemosaic_90k.pth
1x_FatalPak_Beta.pth
1x_FatalTube_Beta.pth
1x_Filmify4k_v2.pth
1x_FrankenMapGenerator-CX-Lite_215000_G.pth
1x_GainRESV3_Aggro.pth
1x_GainRESV3_Natural.pth
1x_GainRESV3_Passive.pth
1x_GameboyCamera_beta#.pth
1x_GameboyCamera_beta.pth
1x_NMKD-FarthestNeighbor_100000_G.pth
1x_NMKD-Jaywreck.pth
1x_NMKD-Jaywreck3-Lite_320k.pth
1x_NMKD-Jaywreck3-Soft-Lite_320k.pth
1x_NormalMapGenerator-CX-Lite_200000_G.pth
1x_Norms2Tex_beta.pth
1x_PixelSharpen_100000.pth
1x_RandComplete_BETA.pth
1x_ReFocus_Cleanly_100000_G.pth
1x_ReFocus_V3_140000_G.pth
1x_Sayajin_DeJpeg.pth
1x_SBDV-DeJPEG-Lite.pth
1x_SBDV-DeJpeg-SuperLite_61000_G.pth
1x_SpongeBC1-Lite_100000_G.pth
1x_SpongeCels-SuperLite_latest_G.pth
1x_SpongeColor-Lite_55000_G.pth
1x_SwatKatsLite_360000_G.pth
1x_test_efix.pth
1x_UnResize_Beta.pth
1x_UnStylize_v3_35000_G.pth
1xBaldrickVHSFix_180000_G_V0.2.pth
1xBC1Smooth2.pth
1xBS_Debandizer.pth
1xBS_DebandizerSPSR.pth
1xBSVectrix.pth
1xBSVectrix24000G.pth
1xCinepakAlt.pth
1xColorizer.pth
1xcrystal.pth
1xDedither.pth
1xDeepfry.pth
1xDeJpegFatalityPlusULTRA.pth
1xDeSharpen.pth
1xDoubleDetoon.pth
1xFatal_Halftone.pth
1xFatalAlias.pth
1xFatalityCinepak.pth
1xFatalityDeBlur.pth
1x-Focus.pth
1x-Focus_Moderate.pth
1xJPEGDestroyer.pth
1xJPG4060Amulet.pth
1xMDeblur.pth
1xMdec.pth
1xmdemotion.pth
1xN64clean.pth
1xNMKD-YandereInpaint.pth
1xNoiseDestructor.pth
1xNothing.pth
1xSB_DDDQ.pth
1xSB_DeditherDequantize.pth
1xSpongeUnResize.pth
1xSSAntiAlias9x.pth
1xSSAntiAliasNT_Aggressive.pth
1xSW1997Dequant.pth
1xSW1997Dequant_Attempt2.pth
1xTexnorm.pth
1xTheMountainV3.pth
1xUnResize.pth
1xZRenoise.pth
2x_Byousoku_5_Centimeter.pth
2x_DigitalFilmV2_Lite8-3_852k.pth
2x_DigitalFilmV4_Lite10-5_420k.pth
2X_DigitalFilmV5_Lite.pth
2x_Faithful-Lite_275000_G.pth
2x_FakeFaith-Lite_105000_G.pth
2x_fidelbd_pokemodel_300000_G.pth
2x_Gen5-Alpha_175000_G.pth
2x_KemonoScale_v2.pth
2x_LD-Anime_Skr_v1.0.pth
2x_MangaScaleV3.pth
2x_NMKD-DeGIF.pth
2x_NMKD-DeGIF_210000_G.pth
2x_NMKD-UpgifLiteV2_210k.pth
2x_SharpAnime_v1.pth
2x_SpongeBob_DVD.pth
2x_SwatKats_154000_G.pth
2x_test_faithful_cx_160000_G.pth
2xBIGOLDIES_415000_G.pth
2xBS_WollyV2.pth
2xBSTexty.pth
2xFaithful32_1316.pth
2xFaithfulSPSR_150000_G.pth
2xMiniRe_125k.pth
2xminiRev2.pth
2x-UniScale_CartoonRestore-lite.pth
4x_1ch-Alpha-Lite_212000_G.pth
4x_AmericanDad2.pth
4x_b_melozard_2_latest_G.pth
4x_BooruGan_600k.pth
4x_BooruGan_650k.pth
4x_BS_DevianceV2.pth
4x_BS_Mystery.pth
4x_BS_Screenbooster_SPSR_88000G.pth
4x_BSRGAN.pth
4x_BSRGAN_old_arch.pth
4x_DeBLRFS.pth
4x_deindeo_x4_130000_G.pth
4x_Deviance.pth
4x_DevianceSPSR_152000_G.pth
4x_DigitalFake_2.1_100000_G.pth
4x_fatal_Anime_160k.pth
4x_fatal_Anime_190k.pth
4x_fatal_Anime_200k.pth
4x_fatal_Anime_260k.pth
4x_fatal_Anime_360k.pth
4x_Fatal_Anime_435k.pth
4x_fatal_Faces_FS_MKI.pth
4x_Fatal_twerpilation.pth
4x_FatalPhotos_Beta.pth
4x_foolhardy_Remacri.pth
4x_foolhardy_Remacri_ExtraSmoother.pth
4x_FS_Dedither_new.pth
4x_GameAI_1.0.pth
4x_GameAI_2.0.pth
4x_GameboyCamera.pth
4x_mdeblur.pth
4x_MeguUp_105000.pth
4x_Minecraft_256hr_55000_G.pth
4x_Minecraft_256hr_spsr_40000_G.pth
4x_NickelbackFS.pth
4x_NMKD_SuperYandere_175k.pth
4x_NMKDfaces.pth
4x_NMKD-PatchySharp_135K.pth
4x_NMKDSuperscale.pth
4x_NMKDSuperscale-Artisoftject.pth
4x_NMKD-Superscale-SP_178000_G.pth
4x_NMKD-UltraYandere_300k.pth
4x_NMKD-UltraYandere-Lite_280k.pth
4x_NMKD-UpgifLiteV2_210k.pth
4x_OLDIES_290000_G_FINAL_interp_03.pth
4x_OLDIES_ALTERNATIVE_FINAL.PTH.pth
4x_Pixelity_Faces.pth
4x_PixelPerfectV4_137000_G.pth
4x_PixelScale_MKII_Beta.pth
4x_PocketMonsters-Alpha_115000_G.pth
4x_reks_effeks_photoanime_v2_150k.pth
4x_RRDB-G_ResNet-D_latest_G.pth
4x_scalenx_90k.pth
4x_Sourcetex-v2-DXTJPG-Smooth_132000_G.pth
4x_Sourcetex-v2-DXTJPG-Smoother_134000_G.pth
4x_SpongeBob-Reloaded_153000_G.pth
4x_SpongeBob-Reloaded-SWAG_153000_swaG.pth
4x_SpongePSNR_85k.pth
4x_Struzan_300000.pth
4x_test_clay_v2.pth
4x_test_Fatality_MK3.pth
4x_test_redither.pth
4x_Training4Melozard_Anime_144000_G.pth
4x_Training4Xeller_latest_G.pth
4x_Training4Xeller_v2_latest_G.pth
4x_trixie.pth
4x_UniversalUpscalerV2-Neutral_115000_swaG.pth
4x_UniversalUpscalerV2-Sharp_101000_G.pth
4x_UniversalUpscalerV2-Sharper_103000_G.pth
4x_Valar_v1.pth
4x-AnimeSharp.pth
4x-AnimeSharp-lite.pth
4xBox.pth
4xBS_DevianceV3.pth
4xCatPatch.pth
4xCharSprite.pth
4xContextualSpongebob_70000_G.pth
4x-DeCompress.pth
4x-DeCompress-Strong.pth
4xDetoon.pth
4xDigiPaint35k.pth
4xdragoon.pth
4xEdges2Faces.pth
4xEdges2Faces_V2Test.pth
4xEdges2Faces_V3Test.pth
4xESRGAN.pth
4x-Fabric.pth
4x-Fabric-Alt.pth
4xFaceB.pth
4xFaceriZor_V1Test.pth
4xFaithful64.pth
4xFalcoon.pth
4xFalloutWeaponsV2.pth
4xFart.pth
4xFatalCats.pth
4xFatalCoon.pth
4xFatalFacesV1.pth
4xFatalFacesV2.pth
4xFatalimiX.pth
4xFatalimiX_MKII_beta.pth
4xFatalimiX_MKIII.pth
4xFatalityV1.pth
4xFatalityV2.pth
4xFatalPixels.pth
4xFireAlpha.pth
4xForest.pth
4xFS_DF2K_jpeg_SDSR.pth
4xFS_DF2K_jpeg_TDSR.pth
4xFSDedither.pth
4xFSDedither_Manga.pth
4xFSDedither_Riven.pth
4xFSDedither_Riven_Smooth.pth
4xFSMangaV2.pth
4xFuzzyBox.pth
4xGround.pth
4xGuilty.pth
4xJaypeg90.pth
4xLADDIER1_282500_G.pth
4xLady0101.pth
4xlollypop.pth
4xManga109.pth
4xMCWashed.pth
4xMeguUp130k.pth
4xMinecraftAlpha.pth
4xMinecraftSPSR_60000_G.pth
4xMinepack.pth
4xMisc.pth
4xNickelback.pth
4xNickelfront.pth
4xPackCraft_v4_40000_G.pth
4xPaper.pth
4xpicsy.pth
4xPixelScale.pth
4xPixelScaleClay.pth
4xportrait.pth
4xQuaker.pth
4xRealSR_DF2K.pth
4xRealSR_DF2K_JPEG.pth
4xRealSR_DPED.pth
4xreboutfs.pth
4xreboutpg.pth
4xRender.pth
4xScreenBooster.pth
4xSGI.pth
4xShiteMountainV3.pth
4xSkyrim_Weapons_And_Armor.pth
4x-SkyrimTexV2.1.pth
4x-SkyrimTexV2_Fabric.pth
4xSmoothRealism.pth
4xsnesal.pth
4xSourcetexV2.pth
4xSourcetexV2_DXTJPG_Smooth.pth
4xSourcetexV2_DXTJPG_Smoother.pth
4xSourcetexV2_Smooth.pth
4xSourcetexV2_Smoother.pth
4xSpongeBob.pth
4xSpongeBobTweak.pth
4xSteveHarvey.pth
4xSW1997.pth
4xSW1997_TEST.pth
4xSW1997Attempt2.pth
4xSW1997Attempt3.pth
4xSW1997V2.pth
4xSW1997V3.pth
4xsynla.pth
4xTest_cout.pth
4xTest_FSBox.pth
4xTest_FSLinear.pth
4xTest_pixy.pth
4xTest_scalegenx.pth
4x-TextSharpV1.pth
4x-UltraSharp.pth
4x-UniScale-Balanced [72000g].pth
4x-UniScaleNR-Balanced [34400g].pth
4x-UniScaleNR-Strong [62400g].pth
4x-UniScale-Strong [42400g].pth
4x-UniScaleV2_Moderate.pth
4x-UniScaleV2_Sharp.pth
4x-UniScaleV2_Soft.pth
4x-VolArt.pth
4x-VolArtNR.pth
4xXplode.pth
4xyammy.pth
4xZZines.pth
8x_BoyMeBob-Redux_200000_G.pth
8x_DONT-USE-THIS-Yikes_50000_G.pth.pth
8x_HugePaint_v1.pth
8x_MC_PrimeHD_17000_G.pth
8x_NMKD-Sphax_150000_G.pth
8x_NMKD-Sphax_200000_G.pth
8x_NMKD-Sphax-DeDither_150000_G.pth
8x_NMKD-Sphax-DeDither_205000_G.pth
8x_NMKD-Superscale_150000_G.pth
8x_NMKD-YanderePixelart4_150000_G.pth
8x_Sphax-Alpha-NN_140000_G.pth
8x_SuperBoyMeBobCX_42000_G.pth
8x_tinyFace_Big_vanilla_patchGan_290000_G.pth
8x_TinyFaceBig.pth
8xNMKD-Typescale.pth
8xrebout.pth
8xSmoothRealismCX_40000_G.pth
8xTGHQFace.pth
CatScale_V0.5.pth
Film-Degrainer_1-000.pth
melozard new model 4.pth
RealESRGAN_x4minus.pth
RealESRGAN_x4plus.pth
RRDB_ESRGAN_x4_old_arch.pth
UniversalUpscalerV2sharp.pth
x2_CGIMaster_v1.pth
x2CGIMaster.pth
Description
Some additions to the original file that I found were omitted:
x1_ITF_SkinDiffDetail_Lite_v1.pth
TGHOQace8x_500k.pth
8xNMKDFaces160000G_v10.pt
8x NMKD-Superscale_150000_G.pth
4xUniscale-Strong [42400g].pth
4x-UniScale-Balanced [72000g].pth
4x-UltraSharp.pth
4x RealisticRescaler_100000_G.pth
1x Usul.v1_465000_G.pth
1x Plants_400000_G.pth
1x Filmify4K.v2_325000_G.pth
FAQ
Comments (11)
I, like the others, appreciate the upload. However, with that many upscalers and having no idea what each upscaler's recommended scaling factor is or the recommended use (realism, anime, etc.) for each, this will be a hard pass for me.
If they were separated based on use (realism, anime, 2.5D, etc.) and scale factor (x4, x8, etc.) that'd be different and your upload would be exponentially more useful to everyone.
As it stands, most don't have the time to blindly dive down a 20GB rabbit hole.
https://openmodeldb.info seems like a better, less time-consuming option.
What a silly comment, ever heard of a large language model: https://chatgpt.com/share/7467bc99-d6dc-467d-9ae0-38a16ad74320
@edmund263 Silly comment? You've only been on this site for two weeks, and you use one of your first comments to create a senseless reply that only served to reinforce my original statement.
So, by your reasoning, we're supposed to download 19GB of data, pay $30 for a ChatGPT Plus subscription (to avoid query limits), sort through 327 files (the majority of which most will likely never use), spend hours asking ChatGPT about each one while compiling its answers into a text file, and then lastly, categorize each of the 327 upscalers based on their individual applications and scale factors?
or...
I can go to OpenModelDB and/or Hugging Face (for free) and spend less than a minute using their search feature to find these, and other upscalers, with zero guesswork. Each upscaler's page is complete with each upscaler's use (anime, realism, etc.), applicable scale factor (2x, 4x, etc.), example images, and other pertinent info to be able to use them effectively. Again, zero guesswork, zero time wasted, and same result.
Now, with the absurdity of your suggestion broken down for you, ask yourself, which of our comments above sounded "silly". Pretty obvious. Just a bit of advice that will help you go a long way on CivitAI...a wise man once said, "Better to remain silent and be thought a fool than to speak and to remove all doubt." A little forethought on your part would have saved you from necroposting a useless reply to a 9 month old comment.
I don't know about you, but some of us have actual jobs and family, and don't have time to spend wastefully on completely foolish and pointless endeavors such as what you are suggesting. If you wish to waste all that time reinventing the wheel compiling that list (as opposed to OpenModelDB), be my guest, but the majority of us have higher priorities and better ways to spend our time.
@Bit_Shifter I provide this for completeness. MAny people like to experiment. IF you dont, then cya later bro, this isnt for you!
@harmonicdiffusion Experimenting and needlessly wasting an incredible amount of time are one and the same in this case, when there are much simpler, time-efficient & headache-free alternatives that are available to everyone...an alternative that produces the EXACT same end result. I can (and have) experiment all day using OpenModelDB without the hours and hours wasted on 19GB worth of literal guesswork, or reliance on ChatGPT to dissect the use and application of each of your 327 random upscalers (as another suggested).
Again, a simple search on two prominent web sites will give anyone free access to hundreds and hundreds of upscalers (including yours), with all the info, scaling factors, applicable use, and example images they need to experiment with them efficiently...and it only takes a moment of their time.
The difference between your uploads and my alternative, is comparable to air-dropping a library's worth of loose random books from an airplane at 10,000 feet with just a single piece of paper containing only their titles, versus, walking into a library with a full card catalog system in place and a librarian standing by to help you out...the same end result, just a lot less work, headache, and wasted time. It's a no-brainer.
Look, I completely understand that you spent a good deal of time compiling these and you're likely to be a bit biased in defending that work. Who wouldn't be? Maybe you didn't know OpenModelDB existed beforehand. Who knows. But no amount of obviously biased comments can defeat the raw logic that says OpenModelDB is FAR more efficient, while producing the EXACT same end result...allowing users to find upscalers they may like based on their chosen application and/or personal preferences.
In any case, that was the only point I was making in my original comment, and it would have been a non-issue today if it hadn't been for someone with exponentially more spare time than common sense replying to my original 9 month old comment with utter nonsense about ChatGPT. That's just turning a headache into a migraine.
"hes an angry elf"
@harmonicdiffusion you might be able to plug it into an AI and have it add all that info in for you?? a coder paired with a text gen idk how that works LOL I'm just chiming in and adding my thoughts. not knowing what they do is half the fun, its like just jumping in your car and driving not knowing where your going just to do it (I actually do this all the time its fun).
You are a God send!
Quick question: which implementation of ESRGAN are these supposed to be used with? The original code from here https://github.com/xinntao/ESRGAN or https://github.com/xinntao/Real-ESRGAN seems to expect a different format.
You sir are crazy... Love you ♡
I am trying to upscale oil painting, watercolor, abstract art and line art. Which models are recommended? Cant test all models