ComfyUI Node
LTEx LTXV Add Per-Window Guide
Inject a guide image at the same relative position in every context window. Requires the LTEx LTXV Context Windows node on the model.
LTEx LTXV Add Per-Window Guide
- positive
- negative
- vae
- latent
- image
- attention_mask
- positive
- negative
◄rel_index-1►
◄strength1.00►
CategoryLTExperiments
Inputs (8)
| Name | Type | Default | Description |
|---|---|---|---|
| positive | CONDITIONING | — | |
| negative | CONDITIONING | — | |
| vae | VAE | VAE used to encode the guide image. | |
| latent | LATENT | The latent being sampled. Used to size the guide encode; not modified. | |
| image | IMAGE | Guide image. Multiple frames fill consecutive positions starting at index. | |
| rel_index | INT | -1-16384–16384 | Position of the guide within every window, counted in latent frames. 0 is the window's first frame; use negative values to place it before the first frame. |
| strength | FLOAT | 1.000–1 | How strongly the guide influences each window. Accepts a single value, or a list of per-window strengths (e.g. from a String to Float List node) indexed by window order, clamping to the last value; set a window's value to 0 to skip the guide for that window. |
| attention_maskopt | MASK | Optional mask limiting where the guide applies. |
Outputs (2)
| Name | Type | Description |
|---|---|---|
| positive | CONDITIONING | — |
| negative | CONDITIONING | — |