Trained several screenshots on Anima V1

Recommended weights: 0.5-0.8 (Feel free to tell me your results)
MUST put negative prompt: 3d, realistic

Right now I think I'll revert to natural language captions. Some objects are very tricky especially body bag.

Not all objects as of yet. Tell me if this work to you. Any advice is appreciated.

---

I am still not satisfied with current version because of captioning method. Still experimenting with steps and caption method because I had no idea how one object able to be generated precisely once but then it fails at other version. Currently re-caption everything slowly.