RM 9.5 and 9.6 Fail to Start

btibert

I am getting a crash report, attached, when I attempt to open 9.5.  As such, I upgraded to 9.6 but am still getting the error.  I can start in safe mode, but this has never happened to me before.  Any thoughts?  Is there a formal place I can submit for technical support?

sgenzer

hi @btibert yes there are numerous bugs with the DL extension. That's why it is in beta We try to keep the Product Feedback page (i.e. bugs) up-to-date as best as we can: https://community.rapidminer.com/categories/product-feedback

That said, Version 0.9.3 was just released this week so please check this and post new bug reports if you find them.

Thanks.

Scott

MartinLiebig

Hi,

this should be somewhat related to deep learning extension. Can you temporarly delete this and check if it works?

Best,

Martin

CC @pschlunder

btibert

Thanks for the note.  What is the best way to do that?  Safe mode?  I am on my other machine at school at the moment, so I will take a look tomorrow.  Thanks, I appreciate it.

MartinLiebig

Hi @btibert,
in your .Rapidminer folder (/home/btilbert/.RapidMiner) you have a folder called extensions. Somewhere there is the deep learning jar (maybe under managed?). Just delete/copy/put it somewhere else and restart studio.

Cheers,
Martin

btibert

Thanks, I actually went ahead and started RM in safe mode, disabled the extension there, and restarted without a crash.  What is bizarre is that I have the DL extension on my other machine, and that does not create an issue.   

Stepping back, is there a bug with the deep learning extension?

btibert

Thanks @sgenzer that did the trick (updating the DL extension).  And thanks for the note on the product feedback area!

pschlunder

Hi @btibert,

glad to hear the update solved your problem. Very likely it was caused due to an incompatibility with macOS Catalina. This was fixed in version 0.9.3. Sorry for the inconvenience.

Please note, that due to a general movement in the market away from providing GPU drivers for Deep Learning libraries on macOS, we also had to drop this support. Hence on macOS we can only support Deep Learning executions on CPU. But still you can build and test your processes on your mac and push them to a RapidMiner Server with GPUs available and make use of them, there.