Krea 2 Tiled Upscale Workflow for Detailed D&D Maps
05032-MendicantBias · reddit · 2026-08-15
A workflow guide for using Krea2 within ComfyUI to upscale and add details to D&D maps. Since standard models struggle to add details during upscaling, this workflow splits the image into tiles. By feeding the original image into CLIP, Krea2's Qwen3VL understands the context and diffuses extra details without requiring manual text prompts for every tile.
The process involves sharpening, splitting the image into 4 tiles, processing each tile with Krea2, and merging them back into a high-resolution, detail-rich image ready for printing.
More from Multimodal
- How meloform works: WASM pitch detection plus differentiable circuits learning 1,022 sine coefficients — henloitsjoyce · 2026-10-16
- Documentary filmmaker urges peers to watch 'You Can See Everything' — kyliebytes · 2026-10-16
- Redditor's first Cumfy/Seedance video nails a cinematic one-shot generation — Wave_N_Wind · 2026-10-16
- Kling 4 returns after 8 months: 10 hyper-realistic AI ads for under $5 — CurieuxExplorer · 2026-10-16
- Shanghai AI Lab unveils Intern Lumina U2, a diffusion LLM unifying image, video and 3D tasks — Shanghai-AI-Laboratory · 2026-10-16
- NAVER AI Lab proposes Speculative Forcing to speed up autoregressive video generation — naver-ai · 2026-10-16