Just a seriously basic workflow for those who have laptops with only 4GB video cards and who still wanna work with Z-Image turbo. Needless to say, with such low specs, you'll need to use extremely quantized models and text encoders. The ones I use:
Text encoders
- Qwen3-4B-Instruct-2507-Q4_K_M
GGUF Model
- Z-image-turbo-q3_k_s
(https://huggingface.co/gguf-org/z-image-gguf/resolve/main/z-image-turbo-q3_k_s.gguf?download=true)
Only uses nodes you can find in the regular manager, nothing fancy. For image to image, make sure to activate the switch in the WF to "True". This is a frankensteined WF, so credits go to their respective authors.
Description
Doubt this will ever be updated, it works as is, but if you find any bugs, please let me know and I'll be happy to fix whatever needs fixing.
FAQ
Comments (1)
Hi, thank you for sharing this workflow! However, I am facing a critical issue. The "Power Lora Loader" node requires the "ZIT-CrimsonNocturne.safetensors" LoRA file, which is completely missing and cannot be found anywhere online.
Furthermore, because of the rgthree node configuration, I cannot bypass the Lora Loader block, and it is locking up the whole pipeline. Even when I try to run a render, the generation process is not working correctly; the colors get completely messed up, and the model fails to follow the prompt or modify the clothing properly.
Could you please upload the missing ZIT-CrimsonNocturne.safetensors file or update the workflow JSON without this dependency? Thanks!