Z-Image Lora Block Explorer - Get more from your LoRAs — v1.0
initial release - requires nodes installed in comfyui
Runtime profile
- Access
- External workflow
- Status
- Listed
- Version
- Z-Image - Multi Image Demo.json
- Source
- Civitai
- Updated
- Oct 18, 2026
Source description
initial release - requires https://github.com/shootthesound/comfyUI-Realtime-Lora nodes installed in comfyui
What you need
- Model files
- 19 GiB known subtotal
Remaining hardware or download-size estimates are not available yet.
Requirements
Models
3GlaraH.safetensors
Not resolvedCheckpoint · Unknown
Text encoder · 7.49 GB · SAFETENSORS · Unknown
z_image_turbo_bf16.safetensors
File specifiedCheckpoint · 11.5 GB · Hugging Face · MonsterMMORPG/Wan_GGUF
Node packs
2ComfyLiterals
FoundNode pack · Registry
comfyui-easy-use
FoundNode pack · Registry
AI-generated commentary
AI-generated explanation based on source and configuration details. Suggestions are clearly labeled.
Purpose
This workflow is for exploring, refining, and rescuing LoRAs with the Realtime LoRA Toolkit. It can take a Z-image LoRA through several block-selection combinations and produce results for those combinations.
Flow
It tries a Z-image LoRA with several block-selection combinations. In the example, the standard "all layers on" approach was among the worst results.
Inputs and outputs
Input: a Z-image LoRA used with selected block combinations.
Output: results for the different block-selection combinations.
Setup
The listed base model is ZImageTurbo.
For setup, search Comfy-Manager for Realtime LoRA Trainer.
Dependencies
The listed node packs are ComfyLiterals and comfyui-easy-use.
The listed model files are GlaraH.safetensors, qwen_3_4b.safetensors, and z_image_turbo_bf16.safetensors.
Configuration
The listed pin for ComfyLiterals is bdddb08ca82d90d75d97b1d437a652e0284a32ac; the listed pin for comfyui-easy-use is 1.3.4.
When exploring a LoRA, try several block-selection combinations.
Suggestion · not verified
Cautions
Before starting, verify that GlaraH.safetensors and qwen_3_4b.safetensors are the intended files.
Suggestion · not verified
Do not assume the standard "all layers on" setting is best; in the example, it was among the worst results.
Suggestion · not verified
Does this need to be edited?
Sign in to send an edit request.
Sources
1 source
Sources
1 sourceModel artifacts
1 artifact
Model artifacts
1 artifactSource excerpts
3 excerpts
Source excerpts
3 excerptsSource context: 287 downloads · Type Workflows · Base model ZImageTurbo
Source context: 280 downloads · Type Workflows · Base model ZImageTurbo
Source context: 251 downloads · Type Workflows · Base model ZImageTurbo
