ResolutionBucket - ComfyUI Built-in Node Documentation

This node organizes a list of latent images and their corresponding conditioning data by their resolution.

This node organizes a list of latent images and their corresponding conditioning data by their resolution. It groups together items that share the same height and width, creating separate batches for each unique resolution. This process is useful for preparing data for efficient training, as it allows models to process multiple items of the same size together.

Inputs

Parameter	Description	Data Type	Required	Range
latents	List of latent dicts to bucket by resolution.	LATENT	Yes	N/A
conditioning	List of conditioning lists (must match latents length).	CONDITIONING	Yes	N/A

Note: The number of items in the latents list must exactly match the number of items in the conditioning list. If the counts do not match, the node raises an error. Each latent dictionary can contain a batch of samples, and the corresponding conditioning list must contain a matching number of conditioning items for that batch. Latent samples may have a shape of (B, C, H, W) for images or (B, T, C, H, W) for videos; the node groups them by height and width only.

Outputs

Output Name	Description	Data Type
latents	List of batched latent dicts, one per resolution bucket.	LATENT
conditioning	List of condition lists, one per resolution bucket.	CONDITIONING

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): 11687f9916895136c7c5b8146cd7519cbf6c296720e453bac52fe4da237403cd