The problem as far as I understand is that libfoo.so's only knowledge about libcuda.so is its SONAME, which is libcuda.so.1, so a binary wanting to link in libfoo.so needs to know who to find libcuda.so.1 in order to verify the available symbols. Generally, ld's suggestion is to use -rpath-link to provide the path to the library, but the nvidia docker repository doesn't come with the symlink to li

low priority issue , it is not a real issue to debug this error , and install bc afterwards , but nevertheless , this is not a native linux command . you should add some warning in , maybe in run.sh for all the diarizations recipe , or in the script itself.

https://github.com/kaldi-asr/kaldi/blob/3e77220b3297831c3a370169580b880461897682/egs/callhome_diarization/v1/diarization/extract_ivecto

PFN CI requires a Google account to browse the test results.
That should be noted at the top of the contribution guide.

Would be great to have new option in Pool. Just like cat_features list of numbers or column names.

Related: #2518

https://github.com/cupy/cupy/blob/1af22f57fda92ae35bde806d0c4d110faf4fed52/cupy/core/reduction.pxi#L47-L49

_J looks unused but actually used in some functions within CuPy.
I'm not sure if it should be documented to the public, but at least there should be a comment.

Is your feature request related to a problem? Please describe.
According to the Arrow spec:

Bitmaps are to be initialized to be all unset at allocation time (this includes padding).

This would imply that bits outside the range [0, size) should always be zero. However, in cuDF/libcudf, we take a more conservative approach and say that bits outside [0,size) are undefined in order to a

Describe the bug
We previously had a test coverage of 82% but now it drops to 75%. Ideally, code coverage should be > 80% for a healthy repo.

To Reproduce
Steps to reproduce the behavior:

1. go to https://codecov.io/gh/uber/aresdb you will see detailed overage for each package, file, method, and lines.

Expected behavior
Ideally, code coverage should be > 80% for a healthy

e.g. based on the google doc https://docs.google.com/document/d/1ZI1V_2I3tETAeGnAwYZnrYTdeiOfdtWCL4l4DmRimFc/

p.s. I think that using docker for the course is an overhead which is not justified. As people create a designated VM for the course it would have been better to share a VM image so installation is one click and forget about docker. Docker is great when you need multiple environment or

Describe the bug
The target clean of the root build.sh script should clean all build artifacts of a build made with the standard names and folder paths. Currently it erases almost everything, but fails to delete the cythonized files of the .pyx source files, like python/cuml/cluster/dbscan.pyx. The script should erase those artifacts as well to trigger a "re-cythonization" of all file

The Python API has a random_seed attribute for the tsnecuda.TSNE class but it's ignored.

Random seed is tracked as an option within the tsnecuda implementation but it looks hard coded to time-based seed instead. Is there a good reason for this?

I might implement and PR this change unless you indicate otherwise or do it first.