Back to Model
K

Krea 2 Turbo Three-Step LoRA

Model Nov 1, 2026

lvladikov's community distillation adapter reduces Krea 2 Turbo sampling from eight steps to three, with stock ComfyUI and diffusers files.

Visit primary link 1 source attached
Author-generated portrait comparison of eight-step Krea 2 Turbo, three-step Turbo without LoRA, and three-step Turbo with the community adapter.
lvladikov's generated comparison: eight-step Turbo, three-step Turbo alone and three-step Turbo with LoRA. Only the two three-step panels share the same noise; the teacher uses its own noise. Author samples, not an AITrending reproduction. Source: lvladikov / Krea 2 Turbo Three-Step LoRA
self-hostedmultimodal

What happened lately

Krea 2 Turbo Three-Step LoRA timeline

The project began October 28. Its November 1 checkpoint 42000 provides matching diffusers and ComfyUI weight formats, a stock-node workflow and an optional MLX workflow using LVNodes. Recommended stock ComfyUI settings are three steps, CFG 1, Euler/simple and strength 1.0.

Speed and visual comparisons are the author’s measurements and generated examples. The adapter can increase grain, double fine edges and change composition relative to eight-step Turbo. It is not an official Krea product; training continues and the Krea 2 Community License Agreement applies.