Skip to main content
The ModelComputeDtype node changes the computational data type (precision) used by a model during processing. It creates a copy of the input model and applies the selected precision setting, which can help optimize memory usage and performance depending on your hardware. This is useful for debugging and testing different precision configurations.

Inputs

ParameterDescriptionData TypeRequiredRange
modelThe input model to modify with a new compute data typeMODELYes-
dtypeThe computational data type to apply to the model (default: “default”)STRINGYes”default"
"fp32"
"fp16"
"bf16”

Outputs

Output NameDescriptionData Type
modelThe modified model with the new compute data type appliedMODEL
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): bc65f1e452d0122ad175a8b95f38a36503253c9908157037c516496e65c828e6