I didn't see any Anima LoRA's for Sylvia so I tried creating one with ChatGPT Codex using GPT 6 Sol High. This is my first Anima LoRA so it might not be perfect. I did two versions, one with danbooru tags and less guidance, and one using the civitAI greg rutkowski Anima lora from circlestone labs as guidance for config and natural language prompting. I'll upload 800, 2000 step versions of the first and 200, 400 step versions of the second. I think the latter has a stronger fit. The Natural Language versions also use Rank 32 while danbooru uses rank 16. Additionally I used <a target="_blank" rel="ugc" href="https://civarchive.com/models/2695493/one-obsessionanima">https://civarchive.com/models/2695493/one-obsessionanima</a> v4 for the sample images.The natural language versions seem to respond to the 3d tag well for matching the game's style so if that's what you're going for I would try those versions of the model. Also a note, the natural language versions will still work fine with danbooru tags, just that the training dataset used natural language prompts. My personal preference is the 400 step natural version. If you try it out I hope you enjoy it!

Sylvia (Potionomics) - Anima LoRA

Files with SHA256: 85f6c8992d1726aa24c1b350e55b308a4b818956a63169bc1b11608b9274ea06