TSR - Temporal Score Rescaling

model_patches/unet

TemporalScoreRescaling

[Post-CFG Function] TSR - Temporal Score Rescaling (2510.01184) Rescaling the model's score or noise to steer the sampling diversity.

Example

JSON Example

{
  "class_type": "TemporalScoreRescaling",
  "inputs": {
    "model": [
      "node_id",
      0
    ],
    "tsr_k": 0.95,
    "tsr_sigma": 1
  }
}

This example shows required inputs only. Connection values like ["node_id", 0] should reference actual node IDs from your workflow.

Inputs

Name	Type	Status	Constraints	Default
model	MODEL	required	-	-
tsr_k?	FLOAT	required	min: 0.01, max: 100, step: 0.001	0.95
tsr_sigma?	FLOAT	required	min: 0.01, max: 100, step: 0.001	1

Outputs

Index	Name	Type	Is List	Connection Reference
0	patched_model	MODEL	No	["{node_id}", 0]

How to connect to these outputs

To connect another node's input to an output from this node, use the connection reference format:

["node_id", output_index]

Where node_id is the ID of this TemporalScoreRescaling node in your workflow, and output_index is the index from the table above.

Example

If this node has ID "5" in your workflow:

• patched_model (MODEL): ["5", 0]

Package: comfyui-extras