Jupyter notebook on Dragonboard


#1

I seems to be relatively easier to install libraries such as opencv in jupyter notebook rather than manually going through each individual line of dependencies to execute on the terminal.
is it advisable to install Jupyter notebook on the Dragonboard/Raspberrypi instead?


#2

hi,

sorry, i haven’t followed very closely the ‘opencv’ threads/discussions here. can someone summarize what the issues are with openCV with our Debian images? is it “just” that Debian only provides an old version of openCV (3.2)?

thanks!


#3

Anyhow a lot of the problems recently were difficulties squeezing the
build of OpenCV onto DB410C. It sounded like the RAM needed to build
might have gone up. Debian testing is currently 3.2 although there is
3.3 in experimental so its possible to use apt pinning to grab a newer
build (although 3.4 is latest version so its still not 100% up to date).

To be honest once OpenCV is built I stop reading as closely ;-). I’m
plenty experienced enough to help with compiling stuff but once its
built then the problems are much less likely to be anything DB410C or
AArch64 related and my experience becomes worth rather less…


#4

It seems to be relatively easier to install libraries such as opencv in
jupyter notebook rather than manually going through each individual
line of dependencies to execute on the terminal. is it advisable to
install Jupyter notebook on the Dragonboard/Raspberrypi instead?

I’m not quite sure what you mean here.

Do you really mean install (as in get the code present in the storage
of the DB410C) or do you mean import (make the opencv symbols available
to the python interpretter).

If the later then yes, it much better to use jupyter than to use the
python interpreter interactively. jupyter helps you record your work
much more easily.