Learning AI, ML, DL, etc.
“Any intelligent fool can make things bigger and more complex...
It takes a touch of genius and a lot of courage to move in the opposite direction.”
- Albert Einstein
- Tutorials and code in this repo, here.
- PyTorch tutorial: https://pytorch.org/tutorials/
- 9 hour video on PyTorch: https://www.youtube.com/watch?v=GIsg-ZUy0MY
- 7 hour video on TensorFlow: https://www.youtube.com/watch?v=tPYj3fFJGjk
Did several tests, specially since RaspiOS 64-bits was released, running TensorFlow - installing it actually - is not always as straightforward as expected.
Trying the Docker route.
See:
Works just fine.
See https://www.tensorflow.org/install/docker for instructions.
$ docker pull tensorflow/tensorflow[:latest-py3]
See the images
$ docker images
REPOSITORY TAG IMAGE ID CREATED SIZE
tensorflow/tensorflow latest-py3 53187075965b 4 months ago 2.52GB
hello-world latest a29f45ccde2a 4 months ago 9.14kB
$
Run a given image
$ docker run -it tensorflow/tensorflow[:latest-py3] bash
________ _______________
___ __/__________________________________ ____/__ /________ __
__ / _ _ \_ __ \_ ___/ __ \_ ___/_ /_ __ /_ __ \_ | /| / /
_ / / __/ / / /(__ )/ /_/ / / _ __/ _ / / /_/ /_ |/ |/ /
/_/ \___//_/ /_//____/ \____//_/ /_/ /_/ \____/____/|__/
WARNING: You are running this container as root, which can cause new files in
mounted volumes to be created as the root user on your host machine.
To avoid this, run the container by specifying your user's userid:
$ docker run -u $(id -u):$(id -g) args...
root@4c8d410e8b57:/# python3
Python 3.6.9 (default, Apr 18 2020, 01:56:04)
[GCC 8.4.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import tensorflow as tf
>>> print("TensorFlow version {}".format(tf.__version__))
TensorFlow version 2.2.0
>>> print("TensorFlow/Keras version {}".format(tf.keras.__version__))
TensorFlow/Keras version 2.3.0-tf
>>> exit()
root@4c8d410e8b57:/#
Connected as above, you can do a
$ python3
>>> help("modules")
. . .
>>> import tensorflow as tf
>>> from tensorflow import keras
>>> from tensorflow.keras import layers
>>> print(tf.__version__)
>>> print(keras.__version__)
There you go!
Do also check
$ docker pull tensorflow/tensorflow:nightly-py3-jupyter
$ docker run -it -p 8888:8888 tensorflow/tensorflow:nightly-py3-jupyter
This will start the container, and a Jupyter notebook server. It comes with cool tutorials and examples.
Reach the server from your browser (port 8888 here), using the URL that shows up in the docker console.