Closed leoheck closed 5 years ago
Also, do you know how can I limit the number of processors being used?
My computer dies when I am using all the cores.
Hi dude! Sorry for my delay.
git clone https://github.com/lhelontra/tensorflow-on-arm
cd tensorflow-on-arm/build_tensorflow/
docker build -t tf-arm -f Dockerfile .
docker run --name tf-arm -v /tmp/tensorflow_pkg/:/tmp/tensorflow_pkg/ -h tf-arm-builder -i -t tf-arm
docker start tf-arm
docker attach tf-arm
cd /root/tensorflow-on-arm/build_tensorflow/
# install nano for edit config target.
apt-get install nano
# for example, we using rpi.conf
pico configs/rpi.conf
# uncomment this lines and change the values:
# BAZEL_AVALIABLE_RAM=1536 # MB
# BAZEL_AVALIABLE_CPU=0.5 # number of cpu cores (1.0 epresenting single full core)
# BAZEL_AVALIABLE_IO=1.0 # workstation I/O capability (with 1.0 representing average workstation)
# save config file (ctrl+x -> y -> enter key) and build.
TF_PYTHON_VERSION=3.5 ./build_tensorflow.sh configs/rpi.conf
Awsome!
Hi dude! (I am Br too haha)
How did you do to generate a tensorflow for python 3.5? I tried to build it here, but the wheel is being created with the name
tensorflow-1.13.0rc2-cp34-none-linux_armv7l.whl
I couldn't find where you set it. Can you point me that?
Also, I am not using your repo to build it, since it is not clear to me (looking the README.me) how to do that.
I am doing the crosscompilation like tensorflow website tells to do.