Open-source openjutsu adds 17 relight presets and two-pass Qwen3-VL scene describing
Devajyoti1231 · reddit · 2026-10-30
The open-source project openjutsu (GitHub: D3voz/openjutsu) shipped an update:
Relight: 17 built-in relighting presets plus custom prompts, powered by Flux Klein.
Image describing: when generating video from a picture, the tool now runs a visual describer (Qwen3-VL 8B by default) to fix the problem where the model only saw motion but not content (a cake becoming an unlabeled shape, candle-blowing expressions lost):
- Pass 1: watches the selected clip at 2 fps, logging actions, held objects, mouth actions, expressions, gaze, other people and camera — deliberately excluding appearance and lighting.
- Pass 2: describes your character image, then rewrites pass 1 to match that character and setting, keeping every action and object while dropping movements the new character can't perform.
- "Relight the video's own person" stops after pass 1.
The text appears in an editable box with a "Describe again" button before rendering.
More from Multimodal
- AI-generated Winamp lighter concept goes viral with retro EQ and playlist details — Delahuntagram · 2026-10-30
- Rapid-H3 generates video + audio in ~2 minutes on a single DGX Spark — vchauhan_ · 2026-10-30
- Midjourney Style Code sref 3594718371 Produces Striking Metallic Armor Fashion Portraits — sergeantsref · 2026-10-30
- Native 4K with MoSS Image: RoPE NTK vs Upscale Workflows Put to the Test — No-Sleep-4069 · 2026-10-30
- SpecSlot Boosts Pixel-Space Image Generation: +2.84 FID on ImageNet-512 at Lower Cost — Jiongfan Yang · 2026-10-30
- PhysInsert from Beihang and Meitu Achieves Physically Plausible Reference-Based Object Insertion — Yue Gong · 2026-10-30