name 'os' is not defined 222

Bash

while working with keras extension, when execute the iris sample getting this error. Any help will be greatly appreciated. I am using Windows 10,  64bit operating system. 

varunm1

RM is currently focusing on a new Deep learning extension based on DL4J

It is already available in extensions. You can check the deep learning extension, it also has GPU support. New changes are coming soon I guess based on discussion with @pschlunder

Here is the link related to that extension:

https://marketplace.rapidminer.com/UpdateServer/faces/product_details.xhtml?productId=rmx_deeplearning

why would I like to do so? I have CoLab with all libraries and better error diagnostics, one can run python code script in Keras, then there are other tools/choices as well. 

Good question, I just run python scripting in RM to do some additional operations on predictions using rapidminer operators and create some visualizations. I can use other tools as well but I am used to this. Another reason in my case is I can have all processes (traditional ML processes built in RM and complex DL in python) related to my project in a single tool rather than two or three software (Anaconda, RM). Just for my convenience

varunm1

Hello @Bash

Looks like it is not finding the "os" module in your python installation. Are you using anaconda?

Did you follow all the steps exactly with the same versions mentioned in the below thread?

https://community.rapidminer.com/discussion/54092/keras-deep-learning-extension

Keras extension is outdated and has lot of issues. You should follow the exact versions specified above thread. RM is currently focusing on a new Deep learning extension based on DL4J. You can also write python scripts using the "Execute Python" operator and run instead of keras extension.

Bash

Looks like it is not finding the "os" module in your python installation. Are you using anaconda? (Yes)

Did you follow all the steps exactly with the same versions mentioned in the below thread? (Yes, but I have anaconda, keras, CNTK, python already an environment buit-up - so i have upgraded versions that I need not to replace for keras extension.)

Keras extension is outdated and has lot of issues. (that's good to know - so better to not spend time on Keras extension)

RM is currently focusing on a new Deep learning extension based on DL4J. (thats good - any expected timeline - i'll be excited?)

You can also write python scripts using the "Execute Python" operator and run instead of keras extension. (why would I like to do so? I have CoLab with all libraries and better error diagnostics, one can run python code script in Keras, then there are other tools/choices as well. Power of RM is its extended configuration using visual process.  thanks for the answer).  

varunm1

RM is currently focusing on a new Deep learning extension based on DL4J

It is already available in extensions. You can check the deep learning extension, it also has GPU support. New changes are coming soon I guess based on discussion with @pschlunder

Here is the link related to that extension:

https://marketplace.rapidminer.com/UpdateServer/faces/product_details.xhtml?productId=rmx_deeplearning

why would I like to do so? I have CoLab with all libraries and better error diagnostics, one can run python code script in Keras, then there are other tools/choices as well. 

Good question, I just run python scripting in RM to do some additional operations on predictions using rapidminer operators and create some visualizations. I can use other tools as well but I am used to this. Another reason in my case is I can have all processes (traditional ML processes built in RM and complex DL in python) related to my project in a single tool rather than two or three software (Anaconda, RM). Just for my convenience