hunyuan_i2v.json

{
    "infer_steps": 20,
    "target_video_length": 33,
    "i2v_resolution": "720p",
    "attention_type": "flash_attn3",
    "seed": 0
}