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 > ComfyUI-PhotoDeblur > Download Deblur Models · Photo Deblur

ComfyUI Node: Download Deblur Models · Photo Deblur

Class Name

PhotoDeblurDownloader

Category
Photo Deblur/download
Author
porfeng (Account age: 756days)
Extension
ComfyUI-PhotoDeblur
Latest Updated
2026-11-11
Github Stars
0.09K

How to Install ComfyUI-PhotoDeblur

Install this extension via the ComfyUI Manager by searching for ComfyUI-PhotoDeblur
  • 1. Click the Manager button in the main menu
  • 2. Select Custom Nodes Manager button
  • 3. Enter ComfyUI-PhotoDeblur 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

Download Deblur Models · Photo Deblur Description

Downloads, verifies, and repairs pretrained photo deblurring model files for ComfyUI.

Download Deblur Models · Photo Deblur:

PhotoDeblurDownloader prepares the pretrained model files needed for photo deblurring in ComfyUI. Displayed as “Download Deblur Models · Photo Deblur,” it checks local model files and manages the download and extraction process through the extension’s weight-management system. It also offers a repair mode that backs up mismatched existing files before replacing them with verified official weights, helping you recover from invalid model files without immediately discarding them. Progress is reported during checking, downloading, extraction, and completion, and local files are checked again on repeated queue operations rather than relying on a cached execution. This node prepares model resources; it does not sharpen or restore images itself.

Download Deblur Models · Photo Deblur Input Parameters:

model

Selects the deblurring model family whose weights you want to prepare. The available options are HI-Diff and MISCFilter, with HI-Diff as the default; numeric minimum and maximum values do not apply. Your selection must match the loader you plan to use afterward: HI-Diff Model Loader or MISCFilter Model Loader. This choice determines which model resources the downloader checks and prepares, not the strength of the deblurring effect. When planning your workflow, note that the supplied MISCFilter loader requires an NVIDIA CUDA GPU, whereas the HI-Diff loader also supports CPU execution; this is a downstream loading requirement rather than a downloader requirement.

variant

Selects the model variant to prepare within the chosen model family. The default is RealBlur_J, and the available options come from the extension’s shared VARIANTS list; its complete definition is not included in the supplied source, so an exhaustive list of choices cannot be confirmed here. Numeric minimum and maximum values do not apply. This setting determines which variant’s weights are checked and downloaded. Select the same variant in the corresponding model loader, because a connected download result is resolved using both the model family and variant. Changing this setting prepares a different set of model resources rather than adjusting a restoration-strength setting.

repair_invalid

Controls whether mismatched existing model files may be backed up and replaced with verified official weights. This is a Boolean setting with the options False and True, and its default is False; numeric minimum and maximum values do not apply. Leave it disabled for routine preparation when you do not intend to authorize replacement of mismatched files. Enable it when existing files are reported as invalid or mismatched and you want the downloader’s repair behavior. The tooltip explicitly states that mismatched existing files are backed up before replacement; the supplied class does not specify the backup naming scheme, backup location, or exact validation messages.

Download Deblur Models · Photo Deblur Output Parameters:

ready

A PHOTO_DEBLUR_READY connection containing the weight-preparation result returned by the downloader. Connect it to the optional download_ready input of the matching HI-Diff or MISCFilter loader to make that loader depend on the downloader and use the prepared file information. This connection is not a loaded neural network or an image: the loader still reads the checkpoint files and constructs the model. Its practical importance is that it links preparation and loading within your workflow, while providing the file information the loader needs for the selected model family and variant.

model_dir

A STRING containing the model-directory path reported by the weight-preparation system. The downloader passes ComfyUI’s configured models directory to that system, but the supplied class does not show the exact subfolder naming convention. Use this output to locate the prepared resources or investigate storage and file-access problems. The same directory path is also displayed in the node’s text feedback. This value is a path, not the model itself, and does not replace the ready connection used by the supplied loaders.

status

A STRING containing the status message returned by the weight-preparation system. It provides human-readable feedback about the preparation result and is also shown in the node’s text feedback alongside the model-directory path. Check this output after execution to understand what the downloader reports about the selected resources. The supplied class does not define a fixed list of possible status strings, so their exact wording and detail depend on the underlying weight-management implementation.

Download Deblur Models · Photo Deblur Usage Tips:

  • Match both model and variant to the loader you intend to use. For example, prepare HI-Diff with RealBlur_J and use the HI-Diff Model Loader with RealBlur_J.
  • Connect ready to the loader’s download_ready input to establish the preparation-before-loading dependency. Simply placing the two nodes on the canvas does not establish this connection.
  • Start with repair_invalid set to False. Enable it when you specifically need to replace mismatched existing files, and allow enough storage for backups as well as replacement files.
  • If your system does not have an NVIDIA CUDA GPU, choose HI-Diff for a workflow using the supplied loaders. Downloading MISCFilter weights does not remove its loader’s CUDA requirement.
  • Expect the node to check local files again each time it executes, including repeated queue operations. This behavior helps detect files that have been removed or changed; it does not by itself mean the weights will be downloaded again every time.
  • Use the progress display to distinguish checking, downloading, and extraction. Preparation can continue after the network transfer finishes because extracting files is a separate stage.
  • Inspect status and model_dir before troubleshooting later restoration stages. This helps distinguish a model-preparation problem from a loader or image-processing problem.
  • You can run this node to prepare weights without connecting a downstream restoration workflow because it is registered as an output node. Connect a matching loader afterward when you are ready to use the model.

Download Deblur Models · Photo Deblur Common Errors and Solutions:

Download fails or connection times out

  • Explanation: The underlying download process may fail if your connection is unavailable, a remote host cannot be reached, or network restrictions block access. This is a troubleshooting description, not a confirmed literal error message: the supplied class does not include the download helper’s exception text.
  • Solution: Check your internet connection, review the ComfyUI console for the actual failing address or network message, and retry. If you use a proxy, firewall, or restricted network, check whether it permits access to the download location.

Existing model files are invalid or mismatched

  • Explanation: The repair tooltip confirms that existing files can be identified as mismatched. Incomplete downloads or files that do not match the expected official weights can require replacement. The exact validation message and behavior with repair disabled are not shown in the supplied class.
  • Solution: If replacement is appropriate, set repair_invalid to True and run the node again. The documented repair behavior backs up mismatched files before replacing them with verified official weights. Keep any intentionally customized weights separately.

Permission denied or model directory is not writable

  • Explanation: Preparation requires access to local storage for downloads, extraction, and potentially backups. A filesystem permission error can prevent those operations. This is a possible system-level error rather than an exception string explicitly defined by this node.
  • Solution: Make sure the account running ComfyUI can write to its configured models directory. Check permissions on the relevant directory and existing files, and use the console’s reported path to identify the location that needs attention.

No space left on device

  • Explanation: Model files, extraction data, and repair backups can exhaust available storage. The supplied source does not specify required disk space, so an exact storage requirement cannot be given.
  • Solution: Free space on the drive containing ComfyUI’s models directory and check any temporary location identified in the error. When repairing files, account for both the backup and the replacement before retrying.

Extraction fails or an archive cannot be read

  • Explanation: Extraction is an explicit preparation stage, and it may fail if downloaded data is incomplete or local filesystem operations fail. The supplied class does not expose the archive format or exact extraction exceptions.
  • Solution: Read the console error to distinguish a damaged download from a storage or permission problem. Resolve the reported cause and retry. If the downloader identifies mismatched local files, use repair_invalid to authorize backed-up replacement.

Loader cannot resolve the connected model files

  • Explanation: The supplied loaders request prepared files for their own model family and selected variant. A mismatch between the downloader’s settings and the loader’s settings can prevent the connected preparation result from supplying the expected files. The exact error text is not provided.
  • Solution: Use the HI-Diff loader for HI-Diff or the MISCFilter loader for MISCFilter, select the same variant in both nodes, and connect ready to download_ready. Run preparation again after correcting the selections.

This MISCFilter adapter requires an NVIDIA CUDA GPU. HI-Diff also supports CPU.

  • Explanation: This exact message comes from the supplied MISCFilter loader, not from PhotoDeblurDownloader. It appears when the loader’s selected processing device is not CUDA, even if the weights were downloaded successfully.
  • Solution: Use an NVIDIA CUDA GPU recognized by ComfyUI, or switch both the downloader and loader to HI-Diff for CPU-supported operation. Re-downloading MISCFilter weights will not resolve a device-compatibility problem.

Download Deblur Models · Photo Deblur Related Nodes

Go back to the extension to check out more related nodes.
ComfyUI-PhotoDeblur
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.

Download Deblur Models · Photo Deblur