Hi,
I’m trying to get the Keras + TensorFlow extension working in KNIME, but I haven’t been successful so far. I’ve been reading blogs, posts, and tried different combinations of Python, TensorFlow, and Keras setups through Conda environments, but I still encounter the same issue.
Here’s the error message I receive:
ERROR Keras Network Learner 4:910 Execute failed:
An error occurred while creating the Keras network from its layer specifications. Details:
An error occurred while trying to launch Python: invalid syntax (<string>, line 3)
The message also suggests that missing Python packages might be the cause:
“In case Python packages are missing, you can create a new Python Conda environment that contains all packages required by the KNIME Deep Learning integrations in the ‘KNIME Deep Learning’ Preferences.”
I tested other example workflows available through the KNIME community, such as “Building Sentiment Predictor – Deep Learning”, and which uses combinations of the Keras Embedding Layer, and the Keras Network Learner node runs without issues there.
Could you please help me identify what might be causing this problem or guide me through the correct setup process?
My current Conda environment is created:
(base) C:\Users\user>conda create -n knime-dl python=3.10
(knime-dl) C:\Users\user>pip install tensorflow==2.13 numpy pandas h5py
(knime-dl) C:\Users\user>pip install scikit-learn
All the installations went through smoothly with no error.
Thank you for your assistance.