LatentUpscaleModelLoader - ComfyUI Built-in Node Documentation

The LatentUpscaleModelLoader node loads a model specialized in upscaling latent representations from a file stored in ComfyUI’s latent_upscale_models folder. It automatically detects the model architecture from the file contents (Hunyuan Video 720p, Hunyuan Video 1080p, or a Latent Upsampler) and configures the matching internal model, so the result is ready to be used by other nodes.

Inputs

Parameter	Description	Data Type	Required	Range
`model_name`	The name of the latent upscale model file to load. The available options are dynamically populated from the files present in ComfyUI’s `latent_upscale_models` directory.	COMBO	Yes	All files in the `latent_upscale_models` folder

Outputs

Output Name	Description	Data Type
`model`	The loaded latent upscale model, configured and ready for use. Depending on the detected file contents, this is a 720p Hunyuan Video upscaler, a 1080p Hunyuan Video upscaler, or a wrapped Latent Upsampler model.	LATENT_UPSCALE_MODEL

This documentation was AI-generated. If you find any errors or have suggestions for improvement, please feel free to contribute! Edit on GitHub

Source fingerprint (SHA-256): 7e23214b1b1fc11be84910a5a209c7990a5199120cb0e6b6c61302a442dcf153

​Inputs

​Outputs

Inputs

Outputs