Harish Kashyap comments

Results 19 comments of


                                            Harish Kashyap

trafficstars

Variable elimination failed when working with large number of nodes

+1 on this. @ankurankan This seems to fail after certain nodes are added. It sometimes does pass. I think its beyond memory error. Steps to reproduce: Take the insurance example...

Variable elimination failed when working with large number of nodes

Oh! I see. Got it. Perhaps I can put up a PR with a message to change variable type? That could help resolve few other issues as well.

Parameter and Structure learning for continuous data

@ankurankan this can be split up into parameter learning and structure learning cards. The parameter learning using MLE can be closed.

Parameter and Structure learning for continuous data

I have added parameter Learnjng for Gaussian Bayesian network

Parameter and Structure learning for continuous data

Not yet. You can still discretize data and use existing methods. Amongst hybrid learning methods a few of them still discretize data for learning.

Parameter and Structure learning for continuous data

Thanks @ankurankan

Stuck after printing 'Successfully opened dynamic library libcublas.so.10.0'

+1. Same issue. It used to work and died without warning on EC2. Later when I reload the pretrained model it hangs with no output.

Stuck after printing 'Successfully opened dynamic library libcublas.so.10.0'

EC2 pytorch AMI tensorflow 2.3

Stuck after printing 'Successfully opened dynamic library libcublas.so.10.0'

No idea. I just used an AMI instance with linux 2 preinstalled with pytorch. Was working fine and now fails to load the pre-trained model.

from lib.model.utils.cython_fbbox import bbox_overlaps

Same issue on ec2 instance.