Skip to main content
The CheckpointLoader node is designed for advanced loading operations, specifically to load model checkpoints along with their configurations. It facilitates the retrieval of model components necessary for initializing and running generative models, including configurations and checkpoints from specified directories.

Inputs

ParameterDescriptionData Type
config_nameSpecifies the name of the configuration file to be used. This is crucial for determining the model’s parameters and settings, affecting the model’s behavior and performance.COMBO[STRING]
ckpt_nameIndicates the name of the checkpoint file to be loaded. This directly influences the state of the model being initialized, impacting its initial weights and biases.COMBO[STRING]

Outputs

ParameterDescriptionData Type
modelRepresents the primary model loaded from the checkpoint, ready for further operations or inference.MODEL
clipProvides the CLIP model component, if available and requested, loaded from the checkpoint.CLIP
vaeDelivers the VAE model component, if available and requested, loaded from the checkpoint.VAE
This documentation was AI-generated. If you find any errors or have suggestions for improvement, please feel free to contribute! Edit on GitHub