BERT model selector

Hi there, I am constantly having issues with BERT model selector even after I configured the python deep learning integration.

ERROR BERT Model Selector  0:171      Execute failed: An error occurred while trying to launch Python: Could not start Python kernel. Error during Python installation test: Library bert is missing.
2021-07-16 14:09:52.438770: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library cudart64_101.dll

screenshots are as shown:



image

I would appreciate any help! Thanks in advance!

Please, can anyone help QAQ?

Hi @zfuak -

Take a quick look at the Python packages required by the Redfield BERT Nodes nodes. You’ll see that you should have bert 2.2.0 in your python environment, which you may need to force using a conda install command at the Anaconda command line.

3 Likes

Thank you so much for your help;)

Sorry for asking again. I have pip installed bert 2.2.0 for the py3_knime_dl_1 environment that I use for bert model selector.


image
But I am still getting this message TAT

ERROR BERT Model Selector  7:171      Execute failed: An error occurred while trying to launch Python: Could not start Python kernel. Error during Python installation test: Library bert is missing.
2021-07-21 11:30:18.575298: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library cudart64_101.dll

I should have been more specific. The Redfield BERT Extension requires Tensorflow2, so you should be adding packages to your other Python environment py3_knime_tf2.

See this blog post for more:

2 Likes

This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.