Get ready:
The l4t 27.1 an Ubuntu 16.04 64-bit variant (aarch64) CUDA 8.0 cudnn-5.1.10 TensorFlow installation requires CUDA and CUDNN installed versions: TensorFlow v1.0. 1
Increase the size of the swap swap area:
Create Script
$ mkdir ~/swap/
$ cd ~/swap/
$ vim createswapfile.sh
//script content as follows
#!/bin/bash
#NVIDIA Jetson TX1 in 3D card
#Create a swapfile to Ubuntu at the current directory location
fallocate-l 4G swapfile
#List out of the file
ls-lh swapfile # Change permissions so ' only
root can use it
chmod swapfile
#List out of the file
ls-lh swapfile #Set up the
Linux swap area
mkswap swapfile
#Now start using the Swapfile
swapon swapfile
#Show that it's now being used swapon-s
//Set permissions
$ chmod +x ile.sh
//execute
$ sudo./createswapfile.sh
//Create swap to 4G size swap file.
//Set power-on effective
$ sudo vim/etc/fstab
//content as follows
/home/ubuntu/swap/swapfile none swap SW 0 0
Installation:
Download script:
$ mkdir ~/tensorflow
$ cd ~/tensorflow
$ git clone https://github.com/ncnynl/ Installtensorflowtx2.git
$ cd installTensorFlowTX2
//Installation dependencies:
//Installation Java,protobuf,grpc-java, Bazel etc
$ chmod +x installprerequisites.sh
$./installprerequisites.sh
//Download TensorFlow code:
$./clonetensorflow.sh
//SET environment variable:
$./settensorflowev.sh
//Compile TensorFlow:
$./buildtensorflow.sh
$./ packagetensorflow.sh
//Installation WHL file:
$ pip Install $HOME/TENSORFLOW-1.0.1-CP27-CP27MU-LINUX_AARCH64.WHL
Test:
$ cd $HOME/tensorflow
$ time Python tensorflow/models/image/mnist/convolutional.py