SaveTrainingDataset - ComfyUI Built-in Node Documentation

Complete documentation for the SaveTrainingDataset node in ComfyUI. Learn its inputs, outputs, parameters and usage.

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

This node saves a prepared training dataset to your computer’s hard drive. It takes encoded data, which includes image latents and their corresponding text conditioning, and organizes them into multiple smaller files called shards for easier management. The node automatically creates a folder in your output directory and saves both the data files and a metadata file describing the dataset.

Inputs

Parameter	Data Type	Required	Range	Description
latents	LATENT	Yes	N/A	List of latent dicts from MakeTrainingDataset.
conditioning	CONDITIONING	Yes	N/A	List of conditioning lists from MakeTrainingDataset.
folder_name	STRING	No	N/A	Name of folder to save dataset (inside output directory). (default: “training_dataset”)
shard_size	INT	No	1 to 100000	Number of samples per shard file. (default: 1000)

Note: The number of items in the latents list must exactly match the number of items in the conditioning list. The node will raise an error if these counts do not match.

Outputs

This node does not produce any output data. Its function is to save files to your disk.