I'm absolutely blown away by this model, so here is a small workflow. Simple 3-Sampler to get a bit more variety out of this model and then a 2x upscale. The resulting images are absolutely incredible.
Laufzeitprofil
Quellenbeschreibung
I'm absolutely blown away by this model, so here is a small workflow. Simple 3-Sampler to get a bit more variety out of this model and then a 2x upscale. The resulting images are absolutely incredible.
The CG-Image Filter node is optional and can be removed. I use it with a batch size of 4 on a RTX 3090 to generate multiple initial images and then filter out the one I like to upscale.
If you generate digital illustrations / drawings, change the prompt in the "Upscale 2x" node.
Custom Nodes:
KI-generierter Kommentar
KI-generierte Erklärung auf Grundlage von Quellen- und Konfigurationsdetails. Vorschläge sind klar gekennzeichnet.
Use Z-Image Turbo - Simple DatAss Workflow with a prompt to generate varied initial images, optionally filter them to choose one, and produce a 2× upscaled output.
It uses a simple 3-Sampler step for more variety, followed by a 2× upscale.
The optional CG-Image Filter can filter the initial images so you can choose one to upscale.
Input: a prompt for generating the initial images.
Outputs include the initial images and the image selected for 2× upscaling.
The workflow uses the ZImageTurbo base model and is published by Datass280.
Named model files are flux_vae.safetensors, qwen_3_4b.safetensors, RealESRGAN_x2plus.pth, and z_image_turbo_bf16.safetensors.
Named node packs are comfyui_ultimatesdupscale (1.6.0), cg-image-filter (1.6.5), and rgthree-comfy (1.0.2511111955).
CG-Image Filter is optional and can be removed.
The description pairs batch size 4 with an RTX 3090 when generating multiple initial images before filtering to one.
If you work with digital illustrations or drawings, consider changing the prompt in the “Upscale 2x” node.
Vorschlag · nicht geprüft
Before using it, verify qwen_3_4b.safetensors and RealESRGAN_x2plus.pth by exact filename.
Vorschlag · nicht geprüft
No license is specified; check the applicable terms before reusing the workflow.
Vorschlag · nicht geprüft
Muss das bearbeitet werden?
Melden Sie sich an, um eine Änderungsanfrage zu senden.
Quellen
1 QuelleQuellenauszüge
1 AuszugSource context: 498 downloads · Type Workflows · Base model ZImageTurbo
UltimateUpscaler: https://github.com/ssitu/ComfyUI_UltimateSDUpscale
CG-Image Filter: https://github.com/chrisgoringe/cg-image-filter
Custom Models:
Real ESRGAN 2x Plus: https://openmodeldb.info/models/2x-realesrgan-x2plus
Geschätzter VRAM Bedarf
Schätzung nicht verfügbar
Die Modell-Dateigrößen sind noch nicht vollständig genug, um diese Anforderung zu berechnen.
Anforderungen
7 AnforderungenUpscaler · Registry · ComfyUI_UltimateSDUpscale
flux_vae.safetensors
Datei ungeprüftVAE · Unknown
RealESRGAN_x2plus.pth
KonfliktUpscaler · Unknown
z_image_turbo_bf16.safetensors
Datei ungeprüftCheckpoint · Unknown
Knotenpaket · Registry
Knotenpaket · Registry