hi all,
I was experimenting with ComfyUI and created a fork called ComfyUI-cpu which is optimized to be used on CPU only devices that do not have any GPUs. ComfyUI-cpu AI text to image using CPU only
use cases :
-- CPU only servers
-- old computers and laptops
** First: download and install ComfyUI-cpu from here : github.com/ArdeniusAI/ComfyUI-cpu
Its recommended for people with no GPU at all in their devices and want to run text to image generation.
-- Make sure to click follow on the top right of this ko-fi page for future updates and releases.
ComfyUI-cpu Model, Lora, and Workflow download includes :
1. SD 1.5 LCM model (2.1 GB) - save this in ComfyUI/models/checkpoints
2. LCM_LoRA_SD15 (135 MB) - save this in ComfyUI/models/loras
3. Workflow SD1.5-LCM.json (10 KB) - open this using load from the UI in ComfyUI
once its running you can replace the sgm-uniform with lcm scheduler for faster generations let me know.
if you have questions in the comments below.
Ardenius