Building on Odroid

[patrickpoirier51]

After a few hours of trial & error, here is my update to the guide:
tensorflow-on-raspberry-pi/GUIDE.md

Step 2 Build Protobuf
git checkout d5fb408d
./autogen.sh : File Name has changed so I copied the latest autogen.sh file from gihub

Step 5. Compiling TensorFlow
git clone: --recurse-submodules https://github.com/tensorflow/tensorflow
git checkout v0.9.0
I was not unable to build bazel 0.3.1 that is required for tensorflow 0.10.0 (current release) so I had to downgrade to release 0.9. with git checkout v0.9.0

Odroid build (error compiling with --copt="-mfpu=neon" on Odroid)
bazel build -c opt --local_resources 1024,2.0,1.0 --verbose_failures tensorflow/tools/pip_package:build_pip_package

Install
bazel-bin/tensorflow/tools/pip_package/build_pip_package /tmp/tensorflow_pkg

I had an error when trying to PIP from /tmp , so I copied file to ~/Downloads
sudo pip install ~/Downloads/tensorflow-0.9.0-cp27-none-linux_armv7l.whl

Hope this can be usefull :-)

[samjabrahams]

Awesome stuff, thanks so much! What model Odroid and OS are you using for this? I don't have the hardware to be able to test this out, but in the meantime I can link to this in the README. Maybe in the future I'll pick up an Odroid!

[patrickpoirier51]

Odroid XU4 - Ubuntu 14.04 (Odroid Image)
Thanks

[alexryan]

I was able to successfully build bazel v0.3.1 and tensorflow v0.10.0rc0 on odroid-c2 by following these instructions [https://www.neotitans.net/install-tensorflow-on-odroid-c2.html]

The author included some very helpful patch files.

It's a little bit different because it's a 64-bit operating system whereas raspbian jessie is 32-bit (even though the processor is 64-bit).

[mohbattharani]

I am trying to compile tensorflow on Odroid XU4 board (OS: FlytOS - https://flytbase.com/odroid-xu4-companion-computer/), but having error shown in snapshots.
FlytOS is ubuntu 16 customized for real time UAV applications. They have already installed ROS and OpenCV on it. I am trying to test some object detection.

I have 16 eMMc for (12 GB taken by OS, 2.4 GB is free) but I have used 16GB USB as swap memory but failed to compile tensorflow (full and lite both). Could it be memory issue or issue with version of bazel (current 0.3.1) and tensorflow? I think Bazel is properly installed

[tjacobs]

I've written up a guide on how to build TensorFlow 1.4.0 on ODROID XU4 Ubuntu 16.04, if that helps anyone:

https://medium.com/@TomPJacobs/running-yolo-on-odroid-yolodroid-5a89481ec141

[mohbattharani]

Thanks. Let me give it a try.

[hntuan94]

Hi mohbattharani,
How about your result?