Skip to content

Conversation

@destinyfaux
Copy link

Z-FUSE: Surgical LoRA Orchestrator A ComfyUI node suite designed for high-fidelity merging and manipulation of LoRAs for the Z-Image (S3-DiT) architecture.

Z-FUSE moves beyond sequential loading, utilizing simultaneous TIES-fusion to prevent "blurry soup" and structural collapse in single-stream diffusion models.

Allows the usage of many Z-image lora nodes in a stack with Lora reporting to view the layers within a Lora model to allow for more precise strength applications that are passed to a merge node that remakes layers and blends the loras discarding noise.

Adding my custom lora nodes for Z-Image. Combines and merges loras non-destructively to enable the usage of many loras without layer saturation.
Fixed indentation issue caused in my prior Z-Image Z-Fuse Node.
@ltdrdata
Copy link
Member

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants