Looks like another great model from your stable, and people would do well to take your prompting advice for any recent model. Even today with excellent models available, most of the text in the typical prompt is either unnecessary due to better models, or never did anything anyway. Far too much of copying of bad prompts and then spreading a probably even worse prompt, rather than starting from nothing.

Is it the same as natural sin one on epicRealism?

Thanks for the great model. I think you are a great creator, so please let me know.

At the moment, do you think it's impossible to create a model that can draw accurate hands and toes even after a lot of training, as long as it's based on SD[1.5,2.1,XL]?

Love your stuff. If you can please upload an inpainting version I will love you even more lol.

Another amazing model, thanks so much for all your hard work! Really loving the unique feel of this one.

it's nice knowing other people out there understand that all of those fake prompt words aren't really doing much for people. It's so much more efficient to keep it simple and let the model do it's part.

Usually when people have problems with limbs, fingers, eyes, etc, it's just because people are overcomplicating their prompt and exceeding vector limits, and then the model has no idea whats going on anymore.

Also, most of those words aren't even in photography model databases, a lot of those came from anime models that pulled that data from danborru tags, so they're mostly irrelevant on the photo side of things.

Focus on words you'd use as if you were a real art director, an art director wouldn't show up to set and say "Hey lets make this shoot lighting look MASTERPIECE, BEST QUALITY, 4K, ABSURDRES"

this guy don't miss

I've been trying out almost every 'realistic' model i could find, but for me personally your models are great! Especially EpiCRealism V5 and now epiCPhotoGasm: But i still need to add a (big) set of negative prompts, and a decent set of positive prompts, without getting disformaties (when prompting more than just: a beautiful women...).

But I use DiffusionBee on a Macbook M1 instead of for example Easy Diffusion. That could be the bottleneck ;-) but if anybody experiences the same problems keep adding the negative prompts,

Wow looks good. Would u upload the checkpoint and negative to tensor.art? I’m not with my machine these few days 😱

Is vae embedded?

bro you're the best, will you do an sdxl version? this is the most realistic model ive found

This checkpoint is just absolutely insane...amazing job

Wow...this model is amazing! Using it without any negative prompt...

wow...who needs sdxl. this is amazing

Really good BUT :

- A lot of lora's seems to not working properly with this checkpoint ( especially Lora model of real people ) even with just a few prompts and also the extension ADETAILER seems to not working :(

the faces and everything all good but only producing the same body type and breast size

With the RC1 and V5, it was bit confusing, so to help you choose, I made a full review here of photogasm with image comparison. If you want to use on colab, check in that video too.

I'm finding this model really good for doing img2img passes on interesting images made with other models.

The best real photo model together with epiCRealism. No other realistic model comes even close.