Pavan Yalamanchili

@umar456 This is fairly similar to `where` but would replace the last step. Instead of writing out the global indices, it will just write out the original values.

@mchandra We are waiting for upstream support for the CUDA backend. If CUDA toolkit 8.0 is released before arrayfire 3.3 (and has eigen value support), we can push that functionality...

@mchandra We try to keep feature parity between all backends. We'll most likely not begin work on CPU and OpenCL backends until the CUDA backend is supported (CUDA being the...

@shehzan10 ?

@shehzan10 any amd gpu will do.

Technically every array inside arrayfire is 4D. `numdims` just counts dimensions until the rest of the dimensions are 1. If you want behavior like numpy, we should implement a wrapper...

@umar456 What do you think about having a separate class for this ?

@umar456 I am not sure if just adding a member function is going to be enough. You will need to add internal objects to hold on to additional metadata that...

@umar456 I was thinking of something in the C++ layer only. This is still a lot of bloat, but we can limit the number of functions that support tensors initially....

Moved the comment to the description.