I am admittedly kind of a noob at this, but if you enjoy my humble work, any donations are greatly appreciated: https://ko-fi.com/insult_to_ninjary
Your generosity helps me stay motivated to keep improving and to keep trying new things!
At a Glance:
Style: Realistic
Works with LoRAs?: It should (I don't use a ton of realistic LoRAs, but it seemed ok!)
VAE: vae-ft-mse-840000-ema-pruned is baked (but you can override if desired)
Clip skip: 1 or 2 (experiment with both!)
Hires. fix: use non-default upscaler; denoising strength 0.5 - 0.4
Version 2 Description:
Since I made one more realistic merge after making the v1 of this "supermerge" of all my realistic models, I decided to re-merge everything again including the new model, for a grand total of 4 realistic merges all smashed together!
While this model is still not my perfect realistic model, it's somehow a lot better than the v1 model, in my opinion. I'm not sure if it's because the new realistic model I merged in had a big impact, or just because merging 4 models together instead of 3 diluted each of them a lot more, but I'm generally pretty happy with how this model turned out, even if I still wish it was a little bit better. And I'm glad that I'm satisfied enough with it, because this is going to be my last realistic checkpoint merge for the foreseeable future.
Please note that while stability and anatomy might be slightly improved from v1, this is still an SD1.5 model, so it's going to have some issues in certain areas. There's a troubleshooting section a bit below here that offers some suggestions if you're having any problems.
For more potentially relevant information, please also see the original v1 description immediately below.
Version 1 Description:
After how successful my Greatest Hits supermerge turned out, I decided to try my hand at a realistic supermerge using all of the realistic merges I've made. I've only made 3 realistic merges tho, so this supermerge is not nearly as "super" as the aforementioned anime one.
I'd say the results of this one are mixed. I originally intended to be a bit harsh on this model here in the description. After I first made it, my initial gens while playing around with it were not great, and the merge somehow seemed less than the sum of its parts. I think all three individual models that went into this merge are arguably each better than the resulting "supermerge."
However, when finally generating the showcase images, they all came together fairly decently with a pretty minimal level of effort, so I do think there's something to be said for ease of use. And someone more skilled at prompting than me also might be able to unlock another level of this thing (I hope!). So I figured it was worth sharing! Please keep in mind that all the standard SD1.5 drawbacks will apply here, most notably questionable anatomy.
Troubleshooting:
If you're having persistent problems with anatomy (hands, fingers, etc.) and/or other jank, just try your best with all the usual fixes - different prompting, use a different VAE, try different clip skip values, use adetailer, inpaint, etc.
Outro:
If you make anything awesome, please consider posting them to the gallery! That goes for all of my models! I really like seeing what kind of stuff you all are able to make!
If you have any questions, feel free to ask! I will try my best to help! Also, if you have any ideas for future models, please don't hesitate to share them with me!