RawSnap v3 LoKr released: retrained from scratch on 749 images, best stacked with v2
Wide_Director_8897 · reddit · 2026-11-07
A Reddit user released RawSnap v3, retrained from scratch as a LoKr (not a standard LoRA) on an expanded dataset of 749 images — 11,235 steps across 15 epochs.
The recommended setup is stacking: RawSnap v3 (epoch 15 or 12) at weight 1.2 combined with RawSnap v2 at 0.35. Available on Civitai.
Related event: RawSnap v3 LoKr retrained on 749 images, stacks with v2(2 posts)→
More from Multimodal
- Suno announces upcoming MCP support, unlocking automated music workflows — TJLarkin23 · 2026-11-08
- DeepSeek V4.1 Flash vs Hy4 tested on voxel pagoda: cleaner vs livelier — Intelligent_Ad2105 · 2026-11-08
- "Serverina" AI Video Delights Reddit: Maltese Dog Working as a Restaurant Server — TradingCardGirl · 2026-11-08
- Open-source Claude Code skill turns a song into a full music video, budget in mind — AIandDesign · 2026-11-08
- AI Time Travel Vlog Takes Viewers to ~1000 AD Lebanon to Witness the Invention of Arak — ResponsibleLoss8149 · 2026-11-08
- Claude + Blender MCP generates a Ghibli-style 3D rooftop scene — sidahuj · 2026-11-08