Save 4 hours! We auto-setup your workflow! Free!

Drop your workflow.json — we handle every dependency, custom node, and model. Just open the link and run.

Auto-Setup Workflow Json (Free) Now!
ComfyUI > Nodes > XB_ToolBox > XB-BOX - VRAM Cleaner

ComfyUI Node: XB-BOX - VRAM Cleaner

Class Name

XTX_VRAM_Cleaner

Category
XB_ToolBox/VRAM_Hacks
Author
WJLUOXIAO (Account age: 324days)
Extension
XB_ToolBox
Latest Updated
2026-05-19
Github Stars
0.04K

How to Install XB_ToolBox

Install this extension via the ComfyUI Manager by searching for XB_ToolBox
  • 1. Click the Manager button in the main menu
  • 2. Select Custom Nodes Manager button
  • 3. Enter XB_ToolBox in the search bar
After installation, click the Restart button to restart ComfyUI. Then, manually refresh your browser to clear the cache and access the updated list of nodes.

Visit ComfyUI Online for ready-to-use ComfyUI environment

  • Free trial available
  • 16GB VRAM to 80GB VRAM GPU machines
  • 400+ preloaded models/nodes
  • Freedom to upload custom models/nodes
  • 200+ ready-to-run workflows
  • 100% private workspace with up to 200GB storage
  • Dedicated Support

Run ComfyUI Online

XB-BOX - VRAM Cleaner Description

Optimizes GPU VRAM usage by cleaning up memory for smoother model execution in AI workflows.

XB-BOX - VRAM Cleaner:

The XTX_VRAM_Cleaner node is designed to optimize the usage of your GPU's Video RAM (VRAM) by performing a thorough cleanup process before generating new data. This node is particularly beneficial for AI artists who work with complex models and large datasets, as it helps to free up VRAM resources, ensuring smoother and more efficient model execution. By unloading all models and clearing caches, the node minimizes memory usage, which can prevent potential out-of-memory errors during intensive tasks. The optional "Nuclear Cleanup" feature provides an even deeper cleaning by emptying the CUDA cache, making it an essential tool for maintaining optimal performance in GPU-intensive workflows.

XB-BOX - VRAM Cleaner Input Parameters:

model

This parameter represents the model that is currently loaded into memory. It is essential for the node to identify which model's resources need to be managed. The model parameter does not have specific minimum or maximum values, as it is dependent on the model being used in your workflow.

positive

The positive conditioning parameter is used to specify the positive conditions or prompts that influence the model's output. This parameter helps in guiding the model towards desired outcomes. It does not have specific value constraints but should be relevant to the task at hand.

negative

The negative conditioning parameter is used to specify the negative conditions or prompts that the model should avoid. This helps in refining the model's output by steering it away from undesired results. Like the positive parameter, it does not have specific value constraints but should be relevant to the task.

latent_image

This parameter represents the latent image data that is being processed. It is crucial for the node to manage the memory associated with these latent representations. The latent image parameter does not have specific value constraints, as it depends on the data being processed.

Nuclear_Cleanup

The Nuclear Cleanup parameter is a boolean option that, when enabled, performs an extensive cleanup of the VRAM by emptying the CUDA cache. This option is particularly useful when dealing with memory-intensive tasks, as it ensures maximum VRAM availability. The default value is True, and it can be set to False if a less aggressive cleanup is preferred.

XB-BOX - VRAM Cleaner Output Parameters:

model

The model output parameter returns the model after the VRAM cleanup process. This ensures that the model is ready for further processing with optimized memory usage.

positive

The positive output parameter returns the positive conditioning data after the cleanup process. This ensures that the conditioning data is preserved and ready for use in subsequent operations.

negative

The negative output parameter returns the negative conditioning data after the cleanup process. This ensures that the conditioning data is preserved and ready for use in subsequent operations.

latent

The latent output parameter returns the latent image data after the cleanup process. This ensures that the latent data is preserved and ready for use in subsequent operations.

XB-BOX - VRAM Cleaner Usage Tips:

  • Enable the Nuclear Cleanup option when working with large models or datasets to ensure maximum VRAM availability and prevent out-of-memory errors.
  • Use this node before starting a new generation process to clear any residual memory usage from previous tasks, ensuring a clean slate for optimal performance.

XB-BOX - VRAM Cleaner Common Errors and Solutions:

CUDA out of memory

  • Explanation: This error occurs when the GPU runs out of available VRAM to allocate for new tasks.
  • Solution: Ensure that the Nuclear Cleanup option is enabled to perform a thorough cleanup of the VRAM. Additionally, consider reducing the size of the models or data being processed.

VRAM cleanup not effective

  • Explanation: This issue may arise if the cleanup process does not free up enough VRAM for subsequent tasks.
  • Solution: Verify that all models are properly unloaded and that the CUDA cache is emptied. If the problem persists, try restarting the application to reset the VRAM usage.

XB-BOX - VRAM Cleaner Related Nodes

Go back to the extension to check out more related nodes.
XB_ToolBox
RunComfy
Copyright 2025 RunComfy. All Rights Reserved.

RunComfy is the premier ComfyUI platform, offering ComfyUI online environment and services, along with ComfyUI workflows featuring stunning visuals. RunComfy also provides AI Models, enabling artists to harness the latest AI tools to create incredible art.

XB-BOX - VRAM Cleaner