Description
An experimental node that patches the WanVideo self-attention mechanism to use a custom sageattention implementation, reducing peak VRAM usage during inference. This node overrides the attention mode of the model.
Requires the latest version of sageattention – ensure it is installed, compatible, and up to date before using this node.
Inputs
| Name | Type | Description |
|---|---|---|
model | MODEL | The WanVideo model to which the memory‑efficient sageattention patch is applied. |
Outputs
| Type | Description |
|---|---|
MODEL | The same model instance with its self‑attention mechanism patched to use sageattention. |
Usage Notes
- This node is experimental – behaviour may change in future releases.
- It only affects the self‑attention layers of WanVideo; cross‑attention or other components remain unchanged.
- Patching is applied as a model wrapper; the output
MODELcan be used directly in subsequent sampling nodes. - If
sageattentionis not installed or is outdated, an error will be raised at runtime. - Running this patch may alter output quality or behaviour – test carefully with your specific workflow.
Comments
Sign in with GitHub to join the discussion.