Krea 2 Turbo with my 4-step LoRA and Nvidia PiD = Super-Fast High-Res (ComfyUI)
retained text
Following up on my [MOSS + PiD earlier post](https://www.reddit.com/r/StableDiffusion/comments/1xjw1jd/moss_image_with_akhaliqs_4step_lora_and_nvidia/), now with focus on Krea 2 Turbo + [Nvidia PiD (Pixel Diffusion Decoder)](https://huggingface.co/nvidia/PiD), with provided workflows and test script and again 100 example images. The workflows allow the choice between running Krea 2 Turbo with either my [4 Step LoRA](https://huggingface.co/lvladikov/Krea2-Turbo-Distill-4step-LoRA) (by default), or my [3 Step LoRA](https://huggingface.co/lvladikov/Krea2-Turbo-Distill-3step-LoRA) or no LoRA for step reduction at all, allowing you to do the full Turbo 8 steps. And with PiD doing the heavy lifting and Krea 2 Turbo rendering at quarter of the output's width and height, the Krea steps go really quickly so you could really afford 4, 5, 6 or even 8 steps if you want to push the quality, and PiD will take care of the high resolution (upscaling straight from the low res Krea 2 Turbo latent). Since Krea 2 does a good job out of the box, this time I have not provided workflows with separate refiner, but the principle still stands and you're welcome to further play as you wish. Full details, review, workflows and examples at my HF: [https://huggingface.co/lvladikov/Krea2Turbo-PiD-SuperFast-HighRes](https://huggingface.co/lvladikov/Krea2Turbo-PiD-SuperFast-HighRes) The 100 examples page, directly: [https://huggingface.co/lvladikov/Krea2Turbo-PiD-SuperFast-HighRes/blob/main/examples/README.md](https://huggingface.co/lvladikov/Krea2Turbo-PiD-SuperFast-HighRes/blob/main/examples/README.md)
document fetched 2026-11-02 00:29 · canonical https://reddit.com/r/StableDiffusion/comments/1xl34bd/krea_2_turbo_with_my_4step_lora_and_nvidia_pid