Workflows and Training Strategies for Photo-to-Vintage-Illustration
xbelanch · reddit · 2026-07-11
A user is seeking technical advice on building a ComfyUI workflow to transform real photos into a specific vintage illustration style (similar to the Choose Your Own Adventure children's book series), with strict preservation of the original composition, pose, and identity.
Since they currently only have scanned illustrations and lack a "photo-illustration" paired dataset, the author lists several potential technical approaches and asks for help:
- Training a standard style LoRA combined with ControlNet or IP-Adapter;
- Fine-tuning image editing models like FLUX Kontext or Qwen Image Edit;
- Training on the more mature SDXL ecosystem;
- Generating a synthetic paired dataset through some method.
More from Multimodal
- GPT Image 2 prompt steers ChatGPT into retro city illustrations — SimplyAnnisa · 2026-07-22
- A browser tool lets you generate Anima images and train LoRAs without setup — DeliciousMoraxMora · 2026-07-22
- Revid auto-generates a video from a watched GitHub repo with no human edits — tibo_maker · 2026-07-22
- Kimi K3 builds a VR companion from one 3D model and a single prompt in 30 minutes — rohanpaul_ai · 2026-07-22
- CapWords Speak launches a photo-based language learning app after 8 months — yihui_indie · 2026-07-22
- Spriterrific tackles the biggest flaw in AI game sprites: broken walk/run cycles — chongdashu · 2026-07-22