Tensorflow/Keras installation conflict

I am new to KNIME going through Codeless Time Series Analysis book. Trying to install
conda install -n py3_knime statsmodels=0.11.1 h5py=2.8 tensorflow-mkl=1.12 keras=2.2.4

and it would not work due to conflicts. I tried it on different versions of Python 3.6 - 3.9

Anyone else experiencing the issue?

Hi TargetFilled,

this specific command does not work on my MacOS either. Why do you need to have the version pinned so specifically? Could you provide us with more details about where you are in the book and what exactly you are trying to do? Did you try your example without pinning the versions and only with the absolutely necessary packages?

Best regards

1 Like

I am following instructions on page 53 (attached) and getting the conflict errors. Running without version also yield conflict (attached notepad)
Python Install Error.txt (210.3 KB)

1 Python Install

@TargetFilled what you can do besides following the KNIME Deep Learning Integration Installation Guide and letting KNIME do the deep learning setup (KNIME Deep Learning Integration Installation Guide) is getting the official KNIME yaml configuration files for deep learning and use them just with miniforge and conda-forge repository to avoid licencing problems and also have a clean and slim setup only settings the minimum restrictions with versions KNIME is recommending:

Or you can try some deep learning setup for Keras and Tensorflow with the environment propagation. I have two prepared for python 3.6 and 3.9 for MacOS and Windows:

I also try to keep this article up to date about KNIME and python in general with additional links for further reference:

It can be a little confusing at first but I would recommend to invest a few maybe hours to grasp the concept of using miniconda or miniforge and conda environments with yaml files and versions. It will greatly enhance your knowledge about how to use python and interact with KNIME. Sometimes the books will struggle somewhat to keep track with all versions and dependencies the latest conda might help.

  • use conda and only conda-forge channel via miniforge or miniconda
  • find the latest KNIME recommended yaml file with a minimum of version restrictions
  • let conda handle all additions and version restrictions and possible updates
  • tell KNIME where to find your python-environment for deep learning
  • destill that into a conda environment propagation node to share and have easy access
  • if you encounter version problems start with only a few restrictions and keep adding and editing them (like @steffen_KNIME has suggested). Check if a KNIME sample workflow would still run

Ok you probably got more than you were bargaining for :blush:.

I plan on writing some more about this with examples and screenshots.