ZEISS Knowledge Base
Help
ZEISS Knowledge Base

Troubleshooting

This chapter contains information on how to tackle common problems.

Fixing a failing ZeissPy installation (Ordinal242/MKL-load-failed error)

In case your ZeissPy installation is failing due to an Ordinal242/MKL-load-failed error, the reason might be that you have some 3rd-Party application installed which modified the System32 folder. During the installation, a warning message will be displayed and inform you about dlls that are in the System32 folder. For a potential solution, take the following steps:

  1. Remove files from the System32 folder.
  2. Set a new environment variable

Removing files from the System32 folder

  1. Open the Windows File Explorer.
  2. Go to C:\Windows\System32.
  3. Search for any MKL- or libiomp related files and rename them by adding a .bak to the filename.
  4. Observe if any other software application breaks.
  5. If an application fails, try to move the DLL files alongside the .exe of the breaking software to another folder. If it works again, you can keep things in the moved state.
  6. Check whether the error is gone and ZeissPy can be installed.

If this does not solve the problem, you can also try Setting a new environment variable.

Setting a new environment variable

  1. In the Windows search, enter Environment Variable.
  2. Click the search entry Edit the system environment variables.
  3. The System Properties dialog opens.
  4. On the Advanced tab, click Environment Variables.
  5. The Environment Variables dialog opens.
  6. Click New.
  7. In the New User Variable dialog, set the Variable name to CONDA_DLL_SEARCH_MODIFICATION_ENABLE and for Variable value enter 1.
  8. Click Ok to close the New User Variable dialog.
  9. Click Ok to close the Environment Variables dialog.
  10. Click Ok to close the System Properties dialog.
  11. Check whether the error is gone and ZeissPy can be installed.

Fixing runtime or compatibility issues for externally trained neural networks

In case you run into import or runtime issues in the context of externally trained neural networks, go to https://pypi.org/project/czmodel/ for information on model and compatibility requirements.

Impressum
Carl-Zeiss-Strasse 22
73447 Oberkochen
Germany
Legal