Nodes/ComfyUI-Tween/Load GIMM-VFI Model
ComfyUI Node

Load GIMM-VFI Model

A ComfyUI node in video/GIMM-VFI with 2 inputs and 1 output.

By ethanfel·Created 6 months ago·Updated 10 days ago· 14
Load GIMM-VFI Model
    • model
    model_pathgimmvfi_r_arb_lpips_fp32.safetensors
    ds_factor1.000
    Categoryvideo/GIMM-VFI

    Inputs (2)

    NameTypeDefaultDescription
    model_pathCOMBOgimmvfi_r_arb_lpips_fp32.safetensorsCheckpoint file from models/gimm-vfi/. Auto-downloads from HuggingFace on first use. RAFT variant (~80MB) or FlowFormer variant (~123MB) auto-detected from filename.
    ds_factorFLOAT1.0000.125–1Downscale factor for internal processing. 1.0 = full resolution. Lower values reduce VRAM usage and speed up inference at the cost of quality. Try 0.5 for 4K inputs.

    Outputs (1)

    NameTypeDescription
    modelGIMM_VFI_MODEL