workflow-templates-broken-zap-2026-06-08/site/knowledge/nodes/custom/EmptyHunyuanLatentVideo.md

EmptyHunyuanLatentVideo

Category: latent/video

Display Name: Empty HunyuanVideo 1.0 Latent

Description

The EmptyHunyuanLatentVideo node is similar to the EmptyLatentImage node. You can consider it as a blank canvas for video generation, where width, height, and length define the properties of the canvas, and the batch size determines the number of canvases to create. This node creates empty canvases ready for subsequent video generation tasks.

Inputs

Parameter	Type	Default	Description
width	Int	848
height	Int	480
length	Int	25
batch_size	Int	1

Outputs

Output	Type
),
],
)

@classmethod
def execute(cls, width, height, length, batch_size=1) -> io.NodeOutput:
    latent = torch.zeros([batch_size, 16, ((length - 1) // 4) + 1, height // 8, width // 8], device=comfy.model_management.intermediate_device() | Latent |

Source: comfy_extras/nodes_hunyuan.py

Used in 4 template(s)