Nodes/HunyuanVideo-1.5 nodes/HunyuanVideo 1.5 Leo Transformer Model Loader
ComfyUI Node

HunyuanVideo 1.5 Leo Transformer Model Loader

A ComfyUI node in HunyuanVideoWrapper1.5 with 6 inputs and 2 outputs.

By yuanyuan-spec·Created 9 months ago·Updated 8 months ago· 27
HunyuanVideo 1.5 Leo Transformer Model Loader
    • HYVID15TRANSFORMER
    • HYVID15TRANSFORMERCONFIG
    path
    resolution480p
    task_typei2v
    transformer_dtypebfloat16
    attn_modeflash
    force_sparse_attnfalse
    CategoryHunyuanVideoWrapper1.5

    Inputs (6)

    NameTypeDefaultDescription
    pathCOMBO1 options: None
    resolutionCOMBO480p2 options: 480p, 720p
    task_typeCOMBOi2v2 options: t2v, i2v
    transformer_dtypeCOMBObfloat169 options: float32, float64, float16, bfloat16, uint8, int8, +3
    attn_modeoptCOMBOflash4 options: flash, flex-block-attn, ptm_sparse_attn, flash3
    force_sparse_attnoptBOOLEANfalseForce to use sparse attention even if the model is not trained with sparse attention.

    Outputs (2)

    NameTypeDescription
    HYVID15TRANSFORMERHYVID15TRANSFORMER
    HYVID15TRANSFORMERCONFIGHYVID15TRANSFORMERCONFIG