Unable to install tensorflow 2.0.0 beta version

I am trying to install tensorflow 2.0.0-beta using command:
sudo pip3 install tensorflow==2.0.0-beta1

but getting the following error:
ERROR: Could not find a version that satisfies the requirement tensorflow==2.0.0-beta1 (from versions: none)
ERROR: No matching distribution found for tensorflow==2.0.0-beta1

Note: I already have tensorflow 1.14.0 installed on the jetson nano developer kit so I don’t think it’s a problem of dependencies.

Hi,

The official TensorFlow doesn’t support ARM platform.

We have built it from source for the Jetson user.
And you can find the installation steps here:
https://docs.nvidia.com/deeplearning/frameworks/install-tf-jetson-platform/index.html

The latest version of our prebuilt is v1.14.0.
If v2.0 is necessary, you can build it from source with the instruction shared in this topic:
https://devtalk.nvidia.com/default/topic/1055131/jetson-agx-xavier/building-tensorflow-1-13-on-jetson-xavier/

Thanks.

Hi,

Did anyone manage to install Tensorflow 2 Beta/RC on Jetson Nano?

I tried following the instructions here as advised in the answer above, but I get the following error:

INFO: From Compiling tensorflow/core/kernels/where_op_gpu_impl_8.cu.cc:
external/com_google_absl/absl/strings/string_view.h(495): warning: expression has no effect

external/com_google_absl/absl/types/optional.h(425): warning: expression has no effect
          detected during instantiation of "const T &absl::optional<T>::operator*() const & [with T=stream_executor::dnn::AlgorithmDesc]" 
./tensorflow/stream_executor/dnn.h(802): here

external/com_google_absl/absl/types/optional.h(425): warning: expression has no effect
          detected during instantiation of "const T &absl::optional<T>::operator*() const & [with T=size_t]" 
./tensorflow/stream_executor/dnn.h(856): here
Killed
ERROR: /home/nano/tensorflow/tensorflow/core/kernels/BUILD:1306:1: output 'tensorflow/core/kernels/_objs/where_op_gpu/where_op_gpu_impl_8.cu.pic.o' was not created
ERROR: /home/nano/tensorflow/tensorflow/core/kernels/BUILD:1306:1: not all outputs were created or valid
Target //tensorflow/tools/pip_package:build_pip_package failed to build
INFO: Elapsed time: 3831.205s, Critical Path: 439.05s
INFO: 2595 processes: 2595 local.
FAILED: Build did NOT complete successfully

If anyone has a hint, please do share.

Cheers

Message “Killed” without further error message hints at memory exhaustion.
Did you create and enable a additional swapfile (with at least 8GB space)?

dkreutz, thanks for your response.
It turns out my swapfile was back to 2G(after a reboot) and that might have caused it. It is currently compiling way beyond where it used to stop and the swapfile usage got beyond 2G at a certain point. I will post again once I have any result.

Cheers

I just wanted to follow up that the installation worked like a charm after increasing the pagefile size.