Kromafine model merge tested: one prompt, painterly style control
bird_frank · reddit · 2026-11-07
The author tested style control with Kromafine (a merge of Kroma 0.3.1 and FinePorn, on Civitai) using the Style Atlas ComfyUI custom node to append style prompts to one simple prompt (a blonde woman in a red dress with a dog at a bus stop). The attached result is a rich Renaissance-oil-painting-style render that follows the prompt closely—a reproducible model + node combo for local image generation.
More from Multimodal
- Someone Made a ComfyUI Anthem by Mixing Local and Cloud AI Models — SyntaxDiffusion · 2026-11-07
- Redditor Argues Image Model Fibo2 Is Severely Underrated After LoRA Tests — Suspicious_Aide2697 · 2026-11-07
- Fully Local AI Cartoon: DGX Spark Renders 22-Second Film in 4.5 Hours, No Cloud — OdinLovis · 2026-11-07
- FlashVSR+Turbo cuts video upscaling time 57% on a single RTX 4090, fully open source — lxfater · 2026-11-07
- Tencent Releases TailorRig Dataset: Tens of Thousands of 3D Mesh Rigging Samples — tencent · 2026-11-07
- Creator turns 2D character designs into 3D with Seedance 2.5 — azed_ai · 2026-11-07