- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hi
I tried to implement a clean installation on raspian stretch in Rpi3 with a 32MB microSDcard.
Applied the usual to setup the latest stretch distro:
sudo apt-get update
sudo apt-get dist-upgrade
Used the Movidius documentation ( and raspian stretch tips) to effect the installation
The following error was returned:
Installation failed: Command 'sudo -H pip3 install –trusted-host files.pythonhosted.org
Cython graphviz scikit-image ' return code=1. Error on line 308' ./install-utilities.sh. will exit
Makefile:47: recipe for target 'install' failed
Any help to correct this error much appreciated
Link Copied
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
That should of course be a 32GB microSDcard!
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
I just recently installed all of this on a Pi and now I wish I wrote it down.
I mistakenly installed version 1 of the NCSDK and version 2 of the NCAppZoo
that did not go well. They need to both be version 2.
That all being said..
I do not recall executing this line sudo apt-get dist-upgrade and everything worked out ok.
I did execute a sudo apt-get update
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Oh and to install the NCSDK version 2 I used
git clone -b ncsdk2 https://github.com/movidius/ncsdk.git
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hi
I took heed of your comment and looked up the 'update' and 'upgrade' mechanisms on stackoverflow.
I think the following is a correct synopsis:
'sudo apt-get update' has the effect of loading newer package lists whereas 'sudo apt-get dist-upgrade' installs the newer package components but does not update to a newer system release. This means 'sudo apt-get update' must always be run before 'sudo apt-get dist-upgrade' .
I think in your case you may have installed against a raspian distro that was not updated to the newest available. In my case I used 2018-04-raspbian-stretch.img and applied the update and upgrade commands.
Result:
Your installation worked.
My installation failed.
I looked into the state of my failed installation and found pip3 was not available ( I think used by python3 needed for the NCSDK installation ) so I applied the following (following on from PINTO's March7 suggestion):
1)
wget https://bootstrap.pypa.io/get-pip.py
sudo python3 get-pip.py
sudo reboot
sudo python2 get-pip.py
sudo reboot
2) Inspection of the Installation error text appeared to imply a problem with Cython. I found that Cython was not installed in the raw distro so I applied:
sudo pip3 install Cython
I think precompiled components are supposed to be loaded if Cython is not present. Perhaps this was the point of failure.
I applied the clone,make,install command again this time it reported 'Installation is complete'
The Rpi distro installed is reported from 'uname -a' as:
Linux raspberrypi 4.14.34-v7+ #110 SMP Mon Apr 16:15 BST 2018 armv71 GNU/Linux
I elected to make the examples with OpenCV installed. So far this has reached 60% on the installation, takes many 10s of minutes!
In summary:
I have tentatively cleared the initial problem by installing pip3 and Cython but I am not clear where the underlying problem lays. Problem in raspian distro or the NCSDK installation? I would appreciate any clarification.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
@jslawton Glad you were able to solve your initial issue by manually installing the dependencies. The OpenCV installation can take a while to build (~4-5 hours or so to build).
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
It took about two days to samples running. I installed Spyder 3 as well because you have break points in python that works sometimes at least and can help you find an issue quicker. Its still not 100% but at least I have some running examples.
- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page