Skip to main content

Patch Sage Attention KJ

KJNodes/experimentalExperimental
PathchSageAttentionKJ

Experimental node for patching attention mode. This doesn't use the model patching system and thus can't be disabled without running the node again with 'disabled' option.

Experimental: This node is experimental and its behavior may change without notice.

Example

JSON Example
{
  "class_type": "PathchSageAttentionKJ",
  "inputs": {
    "model": [
      "node_id",
      0
    ],
    "sage_attention": "disabled"
  }
}

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

Inputs

NameTypeStatusConstraintsDefault
modelMODELrequired--
sage_attention?ENUM
8 options
  • disabled
  • auto
  • sageattn_qk_int8_pv_fp16_cuda
  • sageattn_qk_int8_pv_fp16_triton
  • sageattn_qk_int8_pv_fp8_cuda
  • sageattn_qk_int8_pv_fp8_cuda++
  • sageattn3
  • sageattn3_per_block_mean
required-false
allow_compile?BOOLEANoptional-false

Outputs

IndexNameTypeIs ListConnection Reference
0MODELMODELNo["{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 PathchSageAttentionKJ node in your workflow, and output_index is the index from the table above.

Example

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

  • MODEL (MODEL): ["5", 0]
Was this page helpful?