ComfyUI Node
Test Latent Frame (gradient and noise, no model)
Makes a deterministic test latent from a gradient and seeded noise, blended with the previous latent when given. It runs no model or sampling; steps and cfg only scale the noise.
Test Latent Frame (gradient and noise, no model)
- prev_latent
- current_latent
- next_prev_latent
◄frame_index0►
◄total_frames4►
◄steps20►
◄cfg7.00►
◄seed42►
CategoryParametric/Video
Inputs (6)
| Name | Type | Default | Description |
|---|---|---|---|
| frame_index | INT | 00–10000 | Frame index. From 0 to 10000. Default 0. |
| total_frames | INT | 41–10000 | Total frames. From 1 to 10000. Default 4. |
| steps | INT | 201–150 | Sampling steps. From 1 to 150. Default 20. |
| cfg | FLOAT | 7.000–30 | Guidance scale. From 0 to 30. Default 7.0. |
| seed | INT | 420–4294967295 | Random seed: the same seed repeats the same result. From 0 to 4.29497e+09. Default 42. |
| prev_latentopt | LATENT | Prev latent. |
Outputs (2)
| Name | Type | Description |
|---|---|---|
| current_latent | LATENT | — |
| next_prev_latent | LATENT | — |