Zonkey V7 - Download at your own risk edition.
Trained another slider on base Pony with 1.5MP buckets instead of 1.0MP buckets and applied it to Zonkey 6, . Also applied NTC's Not Simple Background LoRA.
Model likes low CFG.
I found that using score tags in an initial prompt, and performing a restart of most steps using a prompt without them, can produce some nice results, in exchange for slightly less sharpness of detail. Examples in the images with embedded ComfyUI workflows.
Zonkey V6.2 - Hyper 8 step edition
Incorporated the ByteDance Hyper 8 Step LoRA, and NTC's Not Simple Background LoRA, for better backgrounds. I recommend Euler a for first pass, with a CFG of 1.25-1.5, and DPM2++ SDE, or similar for upscaling. Upscaling is recommended. For 2 character scenes, 12 steps can help sort out character details but 8 steps is generally fine for 1 character/pov scenes.
Zonkey V6 - Save the day edition - Now with 50% less sameface
Decided to be less lazy, and do some training with v6. I injected 25% of Pony back into Zonkey v5. I inpainted the faces on 400 photos, 200 close, and 200 far, with this model and used the resulting pairs to make a 2 slider loras, to make things look less pony-ish. The loras were then merged in at different layers.
Zonkey V5 - Baked from scratch edition
Started with all the ingredients of Zonkey V3 and DARE merged them all in random order at 6% each for 50 iterations. Added 12.5% halcyonSDXL_v17 and 5% bemypony_real. Back to original Pony CLIP, so it should train LoRAs better than v4. Brought the brightness back up, for a more neutral style, and reduced the noise from 4.2.
I recommend either not using score tags with this version, running them for an early fraction of steps, or placing them near the end of the prompt. They can help with composition, but they nudge the image towards anime style. 1girl has a similar effect, so leave it out if you don't need it. The hashed token negative I used with previous versions tends to break things now, so I've stopped using it as well.
Zonkey V4.2 - Double Check Your Work edition
The extra noisiness of 4.0 was due to an incorrect setting on one of my CLIP DARE merge nodes. Fixed that, and snuck in a little extra oneFORALLPonyFantasy_v20DPO into the CLIP. 4.2 improves clarity, and a standard number of steps can be used again. All example images(except for the first one) use the same generation data as the 4.0 images, except using 30 steps for both first pass and Hires fix, instead of 40-100. They aren't cherry picked, so there are a few glitches I normally would have left out, but I wanted to show a direct comparison.
Zonkey V4 - Beat to the Punch edition.
Was gonna make a model called Godiva, but the name got scooped. Oh well, this model kicks ass, whatever it's called. Very little character accuracy was sacrificed for the photorealism. The CLIP is now modified. I replaced ~50% ofthe original Pony CLIP with oneFORALLPonyFantasy_v20DPO's CLIP and another 25% with proteusRundiffusionDPO_truereversecubich. Added ~40% proteusRundiffusionDPO_truereversecubich, ~10% datassRev3Pony_rev3, and ~5% damnPonyxlRealistic_damnV20EXTREME to the UNET, V Gradient merged it back with Zonkey V3 and put ~25% enjoyXLSuperRealistic_v30ModifiedVersion in the out layer.
This model is stuffed with creativity. It can be a little noisy, use Euler a, for the Hi-Res fix, instead of a DPM++ sampler if it's too much, I think I found a good balance between noise and creativity. It likes lots of steps, I'd suggest at least 30-40, and Hi-res fix, but overall, the anatomy is more reliable than any previous version, so you shouldn't have as many wasted generations if you give it those extra steps.
Enjoy and get wierd with it. I like seeing all the things that get made with it, whatever it is, so don't be shy about posting.
Zonkey V3 - Mad Surgery edition.
Zonkey V2 got a DARE injection of a CosXL merge of Copax Timeless, Rundiffusion Proteus, Art Universe, Realistic Stock Photo, and People Photography, add difference, paired with DARE removal of Jibmix, AnimeBoys, ToonSphere3D and Animagine. Then it was all DARE merged back into original PonyDiffusion V6 XL.
V3 is more versatile than previous versions, capable of achieving both higher photorealism, and more toon-like styles, but it has to be prompted for. Using the terms "real", or "real life" near the beginning of the prompt, without score tags, can help achieve high levels of photorealism, with better faces, especially when weighted fairly heavily. Using score tags can improve overall aesthetics, but will reduce photorealism, and give faces more anime/cartoon-like shape.
Along with improvements to contrast and color range, Zonkey V3 includes the Blessed VAE to amplify the effect.
Zonkey V2 - Big Hairy Juggs edition.
New Masked DARE injection of Juggernaught X, SDXXXL, Bordello and Ratatoskr. U-Net renormalization relative to PonyDiffusionV6 XL.
Improved photorealism
Better lighting
Sharper detail
Higher background variety and detail
More vivid colors
Fewer artifacts
Better anatomy
Higher character accuracy
Furrier furries
Zonkey was created with the goal of bringing as much photorealism as possible to a Pony model, while attempting to retain its flexibility and prompting power. Like other Pony realism models, it can have problems with eyes when faces are relatively small, but Hires fix often improves them.
All posted images are DPM++ SDE Exponential, 30 steps, CFG 3.5-5.0, Hires Fix Latent(bicubic antialiased), 1.5 or 2x. It can be easier to get better poses with Euler a and DPM++ 2S a, but the detail isn't as high. No ADetailer was used, but it may be helpful in some cases. See posted images for suggested prompting style.
The following checkpoints were used:
Animagine XL v3.1
Art Universe SDXL v2.0
Bordello v1.6
CinematicRedmond v1.0
ChacolRealPonyMixXL (Asian Version) v3.0a
Copax TimeLessXL v11
FULLY_REAL_XL v9.0
Juggernaut XL v8
Pony Diffusion V6 XL
Ratatoskr v3.8
Realistic Freedom Wonderland
RealVisXL v4.0
RunBull_XL v0.4
RsmPornXL v0.81
SDXXXL v3.0
Virile XL v1.0
yudas_woman v3
ZavyChromaXL v6.0
Along with the following Loras:
BoringReality_faces v4
Porn Productivity(multi-concepts) PP-21 v1
RMSDXL Photo XL v1.0
SDXL Offset Example Lora v1.0
Styles for Pony Diffusion V6 XL Photo v2
The Handsomizer v1.0(Pony Diffusion V6)
Description
Zonkey V5 - Baked from scratch edition
Started with all the ingredients of Zonkey V3 and DARE merged them all in random order at 6% each for 50 iterations. Added 12.5% halcyonSDXL_v17 and 5% bemypony_real. Back to original Pony CLIP, so it should train LoRAs better than v4. Brought the brightness back up, for a more neutral style, and reduced the noise from 4.2.
I recommend either not using score tags with this version, running them for an early fraction of steps, or placing them near the end of the prompt. They can help with composition, but they nudge the image towards anime style. 1girl has a similar effect, so leave it out if you don't need it. The hashed token negative I used with previous versions tends to break things now, so I've stopped using it as well.
FAQ
Comments (24)
Amazing model! Thank you for sharing!
Wasn't a fan of 4.2, but 5.0 feels like a big improvement over 3.0
5.0 is great regarding anatomy and gets better backgrounds by default.
I was using 2.0 before and loved its inherent understanding of giantess/macro prompts.
It seems sadly that 5.0 lost that inherent understanding. Even with loras ist hard to get characters bigger than buildings similiar to that zebra in the 2.0 example images :(
Merge them and see how it goes
Best of the real pony models I've tried so far. Nice work!
I love your work babe, and perhaps this is just operator error because I'm a noob and I don't use Comfyui or any of that good stuff... I feel like 5.0 is great as far as generating scenes, anatomy, and ease of use but I can't seem to get that drama and graininess anymore that set your model apart from everyone else's. Though i looked through the pics below and I do see that some people are still getting that grainy look I'm just not sure how to get it there. 馃ゲ
Either way I still think you're model is GOAT but any guidance would be appreciated! 馃檹馃徏 (ELI5 though because most of the time I have no idea what y'all are saying lmao)
iew, ifl, igh, iwj, iwp, ixb, ixe, ixz, jaf, jbm, jfb, jsf, jyk, kmz, ksh, kxg, kzg, lbv, zac, yle, zmj, szw, uiw, vfe, par, pdl, qdl, mbo, mtd, gor, bhz, dit,
Any info on what these style tokens do in your example negatives, or how you came up with them? Most of these are not on the 4ch assembled list.
Thought I found them in the example images of RunbullXL, though I can't find them again. Tried them, and liked them. Though, they often cause freakouts in V5, so I've stopped using them for the most part. AFAIK they all produce ugly results, instead of hashed artist styles.
I've found using these on any pony model improves my image quality, or maybe just meets my aesthetic standards.
V5 is a huge improvement over V4.2, nice work!!
4.2 had its own lovely quirks that I kind of miss.
Every version gets better! V5 is amazing
Any tips on getting better face and eyes? no Adetailer is not working
I actually use dreamshaper xl turbo as a separate adetailer cp to get decent faces. still dialing it in.
Still the top 1 realistic pony models. I've tried a lot of models, but none of them still give that extra-realistic non-human character look
I look forward to seeing more progress with your models. This is the fourth of my favorites; I want to see more.
V5 is my favorite Pony mode, wishing for an eventual v6
Please ping support to have the model enabled for onsite generation
This checkpoint produces some pretty amazing results, but is extremely slow (an order of magnitude or so slower) compared to other checkpoints, and often crashes my a1111 during hires fix. Any idea what I'm doing wrong?
There's nothing unique about the V5 checkpoint file which could cause a performance issue, compared to other Pony/SDXL checkpoints, though V4 and V4.2 have a Custom VAE. It sounds like you're hitting system memory fallback, when your VRAM fills up. You can watch this in the GPU-0 subtab in the Performance tab of the Windows Task Manager. If it fills early in Highres fix, then you should edit your webui-user.bat, and try launching it with various commandline args, like --medvram, --lowvram, and/or --opt-sdp-attention. If it happens at the end of Hires fix, then it's the VAE decode causing it. If you have the command line arg --no-half-vae, it can cause higher usage during VAE decode. Using this VAE usually allows you to disable it. https://huggingface.co/madebyollin/sdxl-vae-fp16-fix/tree/main Also, you can download the Multidiffusion extension, and enable its Tiled VAE functionality, with a tile size that's a resolution your GPU can gen without issue.
i strongly suggest you grab stable diffusion reforged and ditch A1111, i get normal performance with this model on my gtx 1080 with that program.
@wingusblingushahalol聽do A1111 extensions work with Reforged?
@GardenHug聽Generally yes, and it has an extension to search for said compatible plugins.
ZONK