Jetson Nano + Jupyter + Tensorflow

Hello everyone!

I’m a total beginner when it comes to Jetson Nano, so be patient with me :-)

I’m having a problem importing Tensorflow in Jupyter notebook, it gives me this error:
ModuleNotFoundError: No module named ‘tensorflow’
(Using the Python 3 kernel, and Tensorflow is working outside Jupyter)

I have installed Jupyter from this guide: Jetson AI Fundamentals - S1E2 - Hello Camera - YouTube,
and then the latest Tensorflow version from this guide: Installing TensorFlow For Jetson Platform :: NVIDIA Deep Learning Frameworks Documentation

What have I missed or done wrong?

Best regards!

/Jonas

Hi @jonas22, if you are running Jupyter like from that video, that is running inside a container. So it wouldn’t find the TensorFlow that you installed outside of container.

You could instead use the l4t-ml container, which already includes TensorFlow+JupyterLab (along with abunch of other ML libraries) - https://ngc.nvidia.com/catalog/containers/nvidia:l4t-ml

Alternatively, if you don’t want to use container, you would need to install JupyterLab onto your system. You can do that by basically following the same steps as in the container’s Dockerfile here: https://github.com/dusty-nv/jetson-containers/blob/c5303798b61101a9975e909319a5243d50175706/Dockerfile.ml#L153