Rose_Raspberry

Hello again!

This model is the result of merging my previous mix, Black_Bloom, with @FlareAI_Studios's incredible Auralis v0.3. The idea was to preserve the sharpness and predictable behavior from Bloom while adding a bit more wildness and color from Auralis.

And it worked! Please be aware that the model isn't 100% fail-proof due to many nuances. For example, I didn't manage to merge the CLIP embeddings, so this mix uses 100% Bloom's CLIP. It's not the end of the world, but it does come with its own quirks.

Anyway, after some testing, I still like the results — that's why we're here.

As always, my full generation parameters can be found in the descriptions of my pictures, but here's a quick rundown:

First Pass (Base Generation):

• CFG Scale: 5.

• Sampler/Scheduler: I use res_3m and beta57, but standard euler with normal or beta should also work fine if you prefer simplicity.

Upscaling (I use Latent Iterative):

• CFG Scale: 4.

• Sampler/Scheduler: dpmpp_2m with beta worked well for me. I haven't tested other combinations.

It's colorful, it responds well to artistic tags, and it's quite sharp (though, well, not 100% of the time XD). Also, please note that it can react pretty strongly to certain tags — high_contrast, for example. You might want to hit the 'Generate' button a few times, as results can vary significantly between seeds, or play with reducing/increasing the weight of specific tags.

Feel free to experiment! Every time I upload a merge, I think, "It's nothing unique, just ordinary — I can't create anything crazy with it." And then the next day, someone shares their image made with that merge, and it's absolutely stunning! So, if you didn't have much luck on the first try... maybe it's worth another shot? Anything is possible! ^-^

Have a good day! :3