A 5 MB McBess-style LoRA for Krea2 trained on 120 captioned images
Winter_unmuted · reddit · 2026-07-28
A creator shared a McBess-style LoRA for Krea2 and says they finally captured the edgy rubber-hose look they had been chasing for a long time.
- The LoRA is under 5 MB.
- It was trained on nearly 120 meticulously captioned images.
- The source style has a spectrum from complex to simple, and prompts can move across that range.
- They also built a no-caption variant to show that a minimal-effort LoRA still works, just less cleanly.
- With Onetrainer, they claim the full workflow — from cropping/prepping to final model — can take about an hour.
More from Multimodal
- ACE-Step UI claims to generate 4-minute vocal songs locally, free and open source — anselm · 2026-07-28
- A Reddit user builds a Wan 2.2 continuation workflow with LoRAs and ComfyUI — SnooMacaroons1365 · 2026-07-28
- User shares a new Midjourney style with exact prompt settings — azed_ai · 2026-07-28
- Topview launches Film Studio with 3D blocking and micro-expression controls — XFreeze · 2026-07-28
- GaussianGPT uses autoregressive next-token prediction to generate 3D Gaussian scenes — rsasaki0109 · 2026-07-28
- Testing Style LoRAs: How to Isolate Style from Content in Image Generation — Dark_Sytze · 2026-07-28