Hey everyone. I'm sharing my new experimental full finetune of the Anima_Preview1.

For this version, I collected a completely new dataset from scratch - it's entirely different from the one I used for my Flux.1 finetune.

Model Features:

🎛️ Highly Prompt-Sensitive: The stylistic range is quite diverse, but the final output relies heavily on your specific prompting.

📸 Analog & Digital Aesthetics: It can produce a wide variety of looks, from distinct analog grain to "high-quality" digital photos (well, as high-quality as it gets for a 1MP retro resolution).

⚡ Optimized: I've included Q8 and Q6_K_M quants for easier inference.

Honestly, I really love the image quality you can squeeze out of such a small model. However, this is still very much a WIP (Work in Progress).

I would love to hear your feedback and see your generations.