One character across 8 isometric angles: ControlNet, spritesheet LoRA, or image-to-3D rerender?
zimmer550king · reddit · 2026-11-08
A developer making an isometric RPG with locally run models compares three approaches for keeping one character consistent across eight facings with fixed light direction:
- ControlNet pose/canny guidance plus an identity LoRA
- A spritesheet/turnaround LoRA trained on multi-angle references
- Generating one good image, converting it to a rough 3D model, and rendering directions
They ask which works best, how many training images a LoRA needs before face and clothing stop drifting, and whether stitching references into one grid image helps.
Also raised: hiding the "AI look" via fixed palettes, quantizing or small outputs, and which models best handle clothing logic like belts through loops.
More from coding & agent
- A simple prompt coda stops coding agents from constantly asking for your approval — doodlestein · 2026-11-08
- Apple fumbled the OS-level personal agent opportunity, leaving browser-based agents as a workaround — omooretweets · 2026-11-08
- Stanford's Memorilla compresses LLM memory to 16 tokens, 4212x fewer and 27x faster — _AustinCalvert_ · 2026-11-08
- 13 free open-source repos that turn your notes into an agent-ready second brain — nikola_mr64990 · 2026-11-08
- Dev updates Lamport build and mulls fine-tuning a game-dev model — R4nsack · 2026-11-08
- Coinbase CEO Brian Armstrong details his multi-agent workflow with lead agents and one prioritized queue — dair_ai · 2026-11-08