Why is perchance ai image generator is using a new model all of a sudden? and the new model itself is so inferior compared to the old one. Can we manually select the old model to generate images?
You must log in or register to comment.
I would love an option to use the old model!
oh cool that it’s out!
the old model was better than this new model you were able to generate consistent characters now you can’t…it really sucks i spent so much time crafting a prompt that give me exactly what i wanted… so much time wasted.
The new one feels too artsy generic like Dall-e, Midjourney or a dozen others out there. I loved the old one for it’s unique simplicity, I think we lost something special
It’s funny… I’ve spent the past year and a half working on my prompts and fine-tuning my generator(s), for the style/results that I like. I got it to where I was very happy with the results - it was generally nothing resembling a lot of the “generic” images I’ve been seeing elsewhere on the web, and I used a lot of combinations of styles to get there - especially using the artist styles plugin.
Using the same prompts with the new model has generated some cool images, but it definitely seems to fall back to a “samey-ness”, and a definite “default style”.
Hopefully, prompt-tuning and model learning will help. I am having fun playing with it so far.
Please dear dev can’t you allow us to choose the old model manually?
I also want the old model back – I was able to work very well with it, and the new one doesn’t meet my requirements at all… It ignores the style, ignores the anti-description, and creates completely unrealistic, distorted images that constantly “fail.” :-(
Does anyone know of an alternative? Or do we really have to wait until the new models have learned their way around ?
One alternative would be running SD locally, but we are never getting the old model mix… some two years ago someone asked if it was SD 1.5 to which the dev said yes, but that was a long time ago. Also it was said that it uses a different model for each specific genre of prompt and those are basically the top models on civitai, which is likely true as over one year ago the anime model would get that generic look often and the latest one of the old perchance was often sharp and with better hand generation when compared to that.
One other downside of the new model I don’t see anyone talking about is that it cannot replicate faces of known people, which was a trick in the old one to mix famous people to reach a desired face shape. Ironically enough, this is the most needed now, as the new one can only generate people with the average e-girl face.
I would love to know this as well. Surely the old model has to be around someplace else, since this new model seems like dozens others I have seen (and disliked) before I found Perchance.
please let us choose to use the old model. Whether the new one will learn or not, I can’t really do anything with the new one until it learns and I don’t trust that it will ever learn to be the way it was in the old model
the new model is actually better of the older one for me personally,are you guys too used to the old prompts and have truble because of it?
I’ve been scrolling the recent images, and yup, still only seeing e-girls with big ass lips, they all have different prompts, but have the same face hehe
Perchance has downgraded itself
That is what it feels like…feels like it’s been deliberately sabotaged because it was so good.
I just hope it knows more furry characters and furry artists ( contemporary fine art artists would be appreciated too)
dude this new model sucks balls
Yes…the problem isn’t just the slower speed, but also that you have to re-tune the keywords.
With the old one, I had my keywords and always got what I was looking for – now I have to perfect everything again so that the generator produces what I want, and that takes time and is time-consuming.
Hello, I really don’t like the new AI modules! The Ai images look much better before. I hope it will be fixed.
