AnalysisVisual AIJune 26, 2026
User praises Krea2 after LoRA training experiments
A Reddit user trained LoRAs on Krea2 for 3 days, calling it their new favorite model with insane detail in single generations. However, they note composition issues and that LoRAs don't work as well as Qwen.
1 source
More stories today
- Reflect: open-source Claude Code system that learns from corrections
- Fugu-Ultra now works with Claude Code
- AI tool locks ground line on drone videos for real estate
- Testing OpenClaw with 12 subagents for automated QA
- Codex's Sol shows improved intent understanding in QA