Solution to Topaz Photo AI Crashing or Error Loading Models
Introduction
After I upgraded Topaz Photo AI to version 3.4.2, I encountered a frustrating issue where the object removal tool consistently crashed the application, and sometimes other features displayed the "Error Loading Models" error. This guide provides a simple solution to resolve this problem, which involves deleting cached library files used by Topaz Photo AI.
The Solution
This solution involves deleting specific cached library files used by Topaz Photo AI, which are likely related to Apple's Core ML framework. Here's how to do it:
Step 1: Delete the Cached Library Folders
Open Finder.
From the menu bar, select "Go" and then "Go to Folder…" (or use the shortcut
Shift + Command + G
).Paste the following path into the "Go to Folder" dialog and press Enter:
~/Library/Application Support/coreMLCache
Note: This path assumes you are navigating from your user's home directory. The full path on your system is likely
/Users/[YourUsername]/Library/Application Support/coreMLCache
(replace[YourUsername]
with your actual username).Inside the
coreMLCache
directory, you will find one or more folders named "Library."Delete all folders named "Library" within the
coreMLCache
directory.
Step 2: Reopen Topaz Photo AI and Restart (if needed)
- Close Topaz Photo AI completely.
- Reopen Topaz Photo AI.
- Try using the feature that was causing the error (e.g., object removal).
- If the problem persists, restart your computer.
Conclusion
In most cases, deleting these cached library files should resolve the "Model Cannot Be Loaded" error. This forces Topaz Photo AI to rebuild or redownload the necessary model data, potentially fixing any conflicts or corruption that occurred during an update or due to other reasons.
While this solution worked for me on macOS after a specific update, it's worth trying if you encounter this issue regardless of the version you are using or your operating system (although the directory path might be different on Windows).
Let me know in the comments if this solution worked for you or if you have any other tips for resolving this error!