Empty Hunyuan Latent Video

Empty Hunyuan Latent Video Node Overview

The EmptyHunyuanLatentVideo node is similar to the EmptyLatent Image node. You can think of it as providing a blank canvas for video generation, where the width, height and length define the canvas properties, and batch size determines how many canvases to create. This node creates fresh empty canvases ready for subsequent video generation tasks.

Class Information

Class Name: EmptyHunyuanLatentVideo
Category: latent/video
Output Node: True

Empty Hunyuan Latent Video Input Types

Parameter	Comfy Type	Description
`width`	`INT`	Video width, default 848, minimum 16, maximum `nodes.MAX_RESOLUTION`, step size 16.
`height`	`INT`	Video height, default 480, minimum 16, maximum `nodes.MAX_RESOLUTION`, step size 16.
`length`	`INT`	Video length, default 25, minimum 1, maximum `nodes.MAX_RESOLUTION`, step size 4.
`batch_size`	`INT`	Batch size, default 1, minimum 1, maximum 4096.

Empty Hunyuan Latent Video Output Types

Parameter	Comfy Type	Description
`samples`	`LATENT`	Generated latent video samples containing zero tensors, ready for processing and generation tasks.

Empty Hunyuan Latent Video Workflow Example

Tencent Hunyuan Video Model Workflow

Here are some relevant resources curated by ComfyUI Wiki:

Empty Hunyuan Latent Video Node Source Code

ComfyUI Version: v0.3.10
2025-01-07

class EmptyHunyuanLatentVideo:
    @classmethod
    def INPUT_TYPES(s):
        return {"required": { "width": ("INT", {"default": 848, "min": 16, "max": nodes.MAX_RESOLUTION, "step": 16}),
                              "height": ("INT", {"default": 480, "min": 16, "max": nodes.MAX_RESOLUTION, "step": 16}),
                              "length": ("INT", {"default": 25, "min": 1, "max": nodes.MAX_RESOLUTION, "step": 4}),
                              "batch_size": ("INT", {"default": 1, "min": 1, "max": 4096})}}
    RETURN_TYPES = ("LATENT",)
    FUNCTION = "generate"
 
    CATEGORY = "latent/video"
 
    def generate(self, width, height, length, batch_size=1):
        latent = torch.zeros([batch_size, 16, ((length - 1) // 4) + 1, height // 8, width // 8], device=comfy.model_management.intermediate_device())
        return ({"samples":latent}, )