Hello, handsome people!
This is just a simple Z-Image workflow using GGUF models, and Detail Daemon.
I use Qwen3-4B-UD-Q8_K_XL, and z_image_Q8_0, but you can use whatever is best for you.
The links are below. Also, I use res_2s as a sampler, but it's slower than anyone else, so you might want to try others; however, res_2s is a beast!. CFG 3-4 is good. I use 30 steps, but 25 is alright, Alpha 0.5 Beta 0.7 gives good contrast. If you don't have res_2s, you need the RES4LYF node (link below).
Resource links
Text Encoders
Diffusion Models
VAE (flux)
flux_ae.safetensors (A.K.A. ae.safetensor)
GGUF Custom Node
Detail Deamon Custom Node
RES4LYF Custom Node
NOTE: This is a Z-Image workflow, not Z-Image-Turbo, but Civitai doesn't have the category yet. BTW, if you can donate some BUZZ, so I can make LoRAs for Z-Image and Chroma, I would appreciate it.
Description
Sup?
FAQ
Comments (9)
Lets just start using the base part to make it clear. (ZiB, Z-Image-Base) Only think that pointing me that this might be Base is the NOTE hided behind expand that is saying this is not Turbo. (I know this would be a problem until CivitAI add ZiB category)
Ya, I'm actually waiting, because Tongyi's official name for the model is just "Z-Image", so I need to wait until someone decided in Civitai if they will categorize it as Z-Image-Base (ZiB) or just Z-Image (Zi). Edit: Never mind, I just saw Civitai added "Z Image Base", so I fixed it!
Where i can find "res_2s" sampler?
It's part of the RES4LYF custom node pack in ComfyUI. Quick edit, generally if you cant find some strange sampler or scheduler you haven't seen before, its part of that pack.
ComfyUI Manager -> RES4LYF
LoaderGGUF
Error(s) in loading state_dict for NextDiT: size mismatch for x_pad_token: copying a param with shape torch.Size([3840]) from checkpoint, the shape in current model is torch.Size([1, 3840]). size mismatch for cap_pad_token: copying a param with shape torch.Size([3840]) from checkpoint, the shape in current model is torch.Size([1, 3840]).
getting this error with any zimage workflow i use.i have updated everything,still no fix,can someone help me ?
You need to update your ComfyUI, and/or your ComfyUI-GGUF node, and/or probably you're not using the right GGUF file in your GGUF loader.
@gabrielx GGUF loaded from huggingface. GGUF node updated. Comfy updated. Z-image base from huggingface. Same error
@cooperdk no idea, are you using ComfyUI Desktop or Portable? What version?