1.2.1 - Denoise Strong


When i select the “Strong” denoise model it does only run on my CPU and not GPU.

Can you attach you log files for this?

Here it is.

Note: I did switch betweens GPU und CPU while using the software.

2023-02-17-15-44-56.txt (248.6 KB)

Hey Thomas,

It looks like your installation of Topaz Photo AI is missing model files.

Please disable antivirus, firewall, VPN, or proxy that may be on your computer. These are known to block the installer from downloading the model files.

Get a new Topaz Photo AI installer and run it: Topaz Labs Downloads

Let me know if you are still having the issue after.

All models deleted to force a download.

Windows antivirus and firewall disabled but nothing changes.

With TVAI all of this works.

We can manually install the model file to see if that fixes this. Please follow the instructions here.

  1. Click this link to download the missing model file.
  2. Open Topaz Photo AI and go to the logs folder.
  3. Go up one folder and open the models folder.
  4. Move the downloaded model into the models folder.

Let me know if that changes the behavior. If that doesn’t fix it, please send all the log files you have to my Dropbox so I can check for any other errors.

You can securely submit your image(s) to my Dropbox using the link below.
Dropbox File Request

The Model was already there, i did replace it but it didn’t change anything.

Uploaded 2023-03-16-15-27-26.tzlog

Did use the Beta for testing.

Fixed in 1.2.7

Now it did shift back to CPU.


important notice. :upside_down_face:

Its working with RAW files, its not working with tiff files or used as a plugin.

For clarity, is it fixed now and using GPU as expected or is it still running on CPU?

There was an AI Engine update in v1.2.7 which may have fixed that behavior.

2023-03-24-15-00-3.txt (248.3 KB)

No its not fixed.

The Problem is that when i drop in a raw file the GPU is used (here it is ok) but when i drop in a tiff file CPU is used (problem) for Denoise Strong.

Thanks for sending the log. I found the relevant error and shared it with my team.

We will be verifying on a computer at the office and then making a task to fix it.

Unfortunately I do not have a solution at the moment. I can follow up once I have more information.

1 Like