Library keras is not properly installed

Hi knime community!

I am working on time series data and I want to set my deep learning enivrement on knime.

It’s been a week since I got this error and I cannot fix it.
knime

@mohamedEl you might want to take a look at this thread:

Which operating system do you use and which version of KNIME? You might also want to check it the node:

https://hub.knime.com/knime/extensions/org.knime.features.python2/latest/org.knime.python2.nodes.conda.CondaEnvironmentPropagationNodeFactory

And maybe examples from the KNIME Hub using this feature with deep learning in order to see if this can help.

2 Likes

Thank you for your response, I ve looked into what you suggested but nothing works,I am using:
windows 10 pro
knime 4.4.0
I have tried using knime on a ubuntu virtual machine (20.04) and got this error.


I looked into it and it seems that the cpu in my virtual machine does not support avx.

Hello @mohamedEl,

seems you opened up two topics for same issue. Let’s keep discussion in only one. Which one can I close?

The second one:

Br,
Ivan

1 Like

Yes @ipazin, you can close the second one, I apologise.

Thanks,

1 Like

No problem @mohamedEl. Hope you’ll solve your issue.
Ivan

1 Like

Hi @mohamedEl,

Are you trying to use the Keras nodes and TensorFlow nodes, or the TensorFlow 2 nodes?
If you want to use the Keras/TensorFlow nodes, then your configuration shown in the screenshot (“py3_knime_dl_2”) looks good.
If you (also) want to use TensorFlow 2, then please follow the steps mentioned in the error message. That is, press “New Environment…” in the TensorFlow 2 section of the preference page shown in your screenshot, and let KNIME create a suitable environment for you.

I think that should not matter in this regard and is only a warning. Keras/TensorFlow should still work. Or do the nodes really fail because of that? If yes, could you share a screenshot of the error?
Either way, for better performance, I would recommend to use the GPU version of Keras/TensorFlow if possible (that is, in case your virtual machine has access to a GPU).

Marcel

2 Likes

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