Redfield BERT Nodes 1.0.0.202208261324 Does Not Install

On startup I get a banner to install new BERT nodes. I click the banner and do not get he normal “Do you want to restart to install” message. If I restart Knime, the banner appears again. Any ideas?

I tried to do the installation from the Knime update menu and that doesn’t to seem to work either.

Hi @rfeigel -

I just ran into a similar issue myself when updating to 4.6.1 (I already had the BERT nodes previously installed). Upon looking in the .log file in my workspace, I got a message like this:

ERROR: Could not install packages due to an OSError: [Errno 2] No such file or directory: 'C:\\Program Files\\KNIME\\plugins\\se.redfield.bert.channel.bin.win32.x86_64_1.0.0.202208261324\\env\\Lib\\site-packages\\tensorflow\\include\\external\\cudnn_frontend_archive\\_virtual_includes\\cudnn_frontend\\third_party\\cudnn_frontend\\include\\cudnn_frontend_EngineConfig.h'
HINT: This error might have occurred since this system does not have Windows Long Path support enabled. You can find information on how to enable this at https://pip.pypa.io/warnings/enable-long-paths

To fix it in my case, enabling long paths in the Windows registry worked. Are you possibly seeing something similar?

Including @Redfield here in case they were not already aware of this.

Here’s my error log. Makes no sense to me.

Can you check the .log file in your KNIME workspace? Mine is in

C:\Users\ScottFincher\knime-workspace\.metadata

I checked the .log file. There’s nothing recored for this date.

Here’s the log file saved directly from the Knime error log window.
log.txt (247.4 KB)

Its pretty definitely a problem with the BERT installation. Just updated the Vernalis nodes by themselves and it seemed to work fine.

Turns out it’s the same problem I mentioned above. I found this in your log:

ERROR: Could not install packages due to an OSError: [Errno 2] No such file or directory: 'C:\\Program Files\\KNIME\\plugins\\se.redfield.bert.channel.bin.win32.x86_64_1.0.0.202208261324\\env\\Lib\\site-packages\\tensorflow\\include\\external\\cudnn_frontend_archive\\_virtual_includes\\cudnn_frontend\\third_party\\cudnn_frontend\\include\\cudnn_frontend_EngineConfig.h'
HINT: This error might have occurred since this system does not have Windows Long Path support enabled. You can find information on how to enable this at https://pip.pypa.io/warnings/enable-long-paths

You could try resolving this, as I did, using the link given (Maximum Path Length Limitation - Win32 apps | Microsoft Docs) if you have the proper permission on your machine.

I will file a ticket about this from the KNIME side today, since (IMO) if an extension doesn’t properly install, KNIME should handle this more gracefully and give you a more prominent error message.

1 Like

Scott:
I changed the registry key and it worked. Thanks. You’re right, a more elegant solution would be nice.

1 Like