Echo Alpha Conceptualizer
A custom SDXL model I merged and finetuned over the course of several months. Specifically designed to bolster my main model, Echo Alpha, with higher concept knowledge and prompt adherence, I have found merging it into Echo Alpha to be a long and frustrating process that after several weeks of on-and-off merging attempts I put on the backburner while I focus on real work outside of model development. When used as a merging resource, it is advised to use it at low strengths as its effects are strong and come at the cost of image quality. From my limited testing, I have found that it understands most, if not all, of common consumptive material. If I were to make a model comparison, this model is closer to my SD 1.5 Alpha Tango model than anything else on-site. Using 'realistic', 'photorealistic', or 'real e621' tags is not recommended when prompting with this model, especially if it's with Pokémon, Digimon, or Anime, unless you like the SD 1.5 pseudo-realism look like Alpha Tango has.
Usage Tips:
Steps: 15-50
Sampler: whatever one you want, I would recommend DPM++ 2M, but that's just my bias. Use the sampler you find works best for the model.
CFG: 3 - 12, adjust to each model, 7.5 is fine but mileage may vary.
VAE: None is fine, adding one is no problem though.
Blah, blah, blah, you get the point. If you're using SDXL, you know what to do. These aren't natural language models and I can not attest to the models' capabilities while using natural language prompting.
Do with these models as you please, crediting unnecessary.
Feel free to post whatever images created in the gallery below.
I'm not responsible for generations produced with these models, generate responsibly.
