Moss Struggles at Character LoRA Training While Krea 2 Nailed It in 2500 Steps
chakalakasp · reddit · 2026-10-21
A Reddit user reports that while Krea 2 was trivially easy to train for likeness reproduction — nailing his face in about 2500 steps — the newer Moss model falls far short even when overtrained past 6000 steps, whether using aitoolkit on RunPod or fizgig locally. He's asking the community for working settings, highlighting a notable gap between new models on character-consistency fine-tuning.
More from Multimodal
- Controlling CapCut Web with Codex: a natural-language workflow for K-pop fan edits — SarahAnnabels · 2026-10-21
- 8 GitHub repos for making videos with Claude Opus 5.5, top collection at 7.2k stars — yihui_indie · 2026-10-21
- Gemini 4 Argon Builds a Pettable 3D Spider with 65,000 Fur Strands and Full Leg Physics — minchoi · 2026-10-21
- First Try: Gemini 4 Argon Recreates an Entire Video as Animated SVG in 20 Minutes — minchoi · 2026-10-21
- Gemini 4 Argon One-Shots Six Interactive Single-File 3D Demos with 3D-to-SVG Projection — minchoi · 2026-10-21
- Pinokio app wraps Fish Audio's new Drama 3 TTS, fixing silent takes and missing words — Thomas_Eric · 2026-10-21