🚀 Launching a New Chapter in My Project!

In Phase 2, I'll build on everything I learned during Stage 1 to create a hybrid model using the multi-view character images generated earlier. The goal? To eliminate reliance on Charturnerv2 while enhancing creative consistency through strategic angle variation.

I’ll also refine the dataset by adding manual tags that specify each character’s class, occupation, or role. This metadata layer will improve model accuracy and narrative depth, ensuring outputs align better with creative visions.

The outcome is still experimental—but that’s the thrill of innovation! I’d deeply appreciate tips or advice from the community, especially since I’m currently unable to run SDXL locally and rely on platforms like Civitai for testing.

I’ll keep you all posted on my progress, and I wish everyone success in their own creative journeys!