ComfyUI Node
VFI Optimizer
A ComfyUI node in video/Tween with 4 inputs and 2 outputs.
VFI Optimizer
- images
- model
- images
- settings
◄min_free_vram_gb2.0►
◄force_batch_size0►
Categoryvideo/Tween
Inputs (4)
| Name | Type | Default | Description |
|---|---|---|---|
| images | IMAGE | Input images — only the first 2 frames are used for calibration. | |
| model | * | Any pairwise VFI model (BIM, EMA, SGM, GIMM, SPEED). LDF-VFI uses a separate sequence pipeline. | |
| min_free_vram_gb | FLOAT | 2.00–48 | VRAM to keep free for other tasks (ComfyUI, OS, etc). Higher = safer but slower. |
| force_batch_sizeopt | INT | 00–64 | Override auto-computed batch_size. 0 = auto. |
Outputs (2)
| Name | Type | Description |
|---|---|---|
| images | IMAGE | — |
| settings | VFI_SETTINGS | — |