I’m trying to set-up a deep-learning pipeline starting with the classic example of the “Cats and Dogs” and following all the different steps indicated in the Deep Learning Integration Installation guide.
I am almost there, but there is one thing that keeps on failing: the installation of the keras library.
If you have any thoughts, please let me know!
Itlooks like the Keras library is not installed correctly, and you may need to downgrade the TensorFlow version in your current Python environment. Check the post below for possible solutions to resolve the issue.
Hi @k10shetty, thanks for answering my post and apologize for my late answer. I have been busy trying everything I believe was possible to install Keras properly but it kept on failing. To give you a bit of context, I first tried to set it up on my personal computer at home. It took me around 1h and everything was running nice and smooth, all nodes were green and I generated the model without issue. Now I’m trying to set that up on one of our VM at work, and I’ve been struggling for about a month, always failing with a different issue, but most of the time it comes back to “Kyras library not being installed properly”. I even exported my conda env from my personal computer (on which it is working), imported them on the VM at work, everything was set exactly the same, and still it failed. Do you know if some security restriction / admin rights / any kind of layer set up by IT infrastructure can block the installation or the communication between Knime and anaconda?