CheckpointLoader - ComfyUI Built-in Node Documentation - ComfyUI

The CheckpointLoader node loads a pre-trained model checkpoint together with its configuration file. It takes a configuration file and a checkpoint file as inputs and returns the loaded model components — the main model, CLIP model, and VAE model — for use in the workflow. This node is deprecated.

Inputs

Parameter	Description	Data Type	Required	Range
`config_name`	The configuration file that defines the model architecture and settings	COMBO	Yes	Available config files
`ckpt_name`	The checkpoint file containing the trained model weights and parameters	COMBO	Yes	Available checkpoint files

Note: This node requires both a configuration file and a checkpoint file to be selected. The configuration file must match the architecture of the checkpoint file being loaded.

Outputs

Output Name	Description	Data Type
`MODEL`	The loaded main model component ready for inference	MODEL
`CLIP`	The loaded CLIP model component for text encoding	CLIP
`VAE`	The loaded VAE model component for image encoding and decoding	VAE

Important Note: This node has been marked as deprecated and may be removed in future versions. Consider using alternative loading nodes for new workflows.

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): 820cd9f7a5ccd5a70d2b29906c8deca3632d2ccba84ca51022717e061afb72b3