The Altair Community is migrating to a new platform to provide a better experience for you. In preparation for the migration, the Altair Community is on read-only mode from October 28 - November 6, 2024. Technical support via cases will continue to work as is. For any urgent requests from Students/Faculty members, please submit the form linked here
GraphViz failing while running Keras examples
Hi Forum
Trying to run Keras extension. I have installed all relevant libraries using native pip - python 3.5.2 (and also in anaconda, using conda install - python 3.6.6). I have tensorflow, keras, pydot and graphviz all installed. However when i execute a Keras example model, RM errors out with the attached error message.
All PATH attributes appear ok. please see below.
any insights? Thanks.
0
Answers
Hi @bdeshpan,
I encountered a similar problem with Keras extension.
I had to update a component, but unfortunately I don' t rememeber its name (it seems this is "flasq" or "flask"...)
The screenshots you shared are not visible in your post.
Can you try again by using "Photos" :
With the name and description of your error, i will try to recover this component.
Regards,
Lionel
Hi again @bdeshpan,
I remember now : You have to update the component "dask".
Open a Windows CMD and execute the following command :
I hope it helps,
Regards,
Lionel
Hi @lionelderkrikor
Thanks for your help. Unfortunately after installing dask like you suggested, it still gives the same error that i captured in the screenshot in my original post.
**********************************************
*********************************
Hi @bdeshpan,
Can you perform the test of Keras installation ?
Go to Settings -> Preferences -> Keras and click on the "test" button :
Regards,
Lionel
I went though this yesterday to install Keras into beta 9. Hitting test will show that everything is installed properly but the process will fail. I solved the problem by downloading Graphviz and setting the environment variable to the bin directory. Reinstall graphviz in python but use pip and not conda. Make sure pip is updated. Type echo %PATH% to be certain that your paths are in order. In Rapidminer, make sure you are pointing to the correct python environment if you have set one up.
I have python 3.6.6 / Tensorflow 1.9- GPU / Keras 2.2.2 / Graphviz 3.8 and everything is working in beta 9 on Windows 10 but I did have to mess with it a bit. It was not the first time that I have installed this but the first time that I am using the GPU under windows.
Let me know if you have any further questions or if you have already solved the problem.
regards,
Alex
Conda list shows the following: