openeuler/tensorflow repository overview

⁠Quick reference

• The official TensorFlow docker image.

• Maintained by: openEuler CloudNative SIG⁠.

• Where to get help: openEuler CloudNative SIG⁠, openEuler⁠.

⁠TensorFlow | openEuler

Current TensorFlow docker images are built on the openEuler⁠. This repository is free to use and exempted from per-user rate limits.

TensorFlow is an end-to-end open source platform for machine learning. It has a comprehensive, flexible ecosystem of tools, libraries, and community resources that lets researchers push the state-of-the-art in ML and developers easily build and deploy ML-powered applications.

Read more on TensorFlow Website⁠.

⁠Supported tags and respective Dockerfile links

The tag of each tensorflow docker image is consist of the version of tensorflow and the version of basic image. The details are as follows

⁠Usage

In this usage, users can select the corresponding {Tag} based on their requirements.

• Pull the openeuler/tensorflow image from docker

  docker pull openeuler/tensorflow:{Tag}
  

  Copy

• Run with an interactive shell

  You can also start the container with an interactive shell to use tensorflow.

  docker run -it --rm openeuler/tensorflow:{Tag} bash
  

  Copy

• Introduction to Tensorflow with MNIST Example

  Create a python file named mnist_example.py with the following content:

  import tensorflow as tf
  mnist = tf.keras.datasets.mnist
  
  (x_train, y_train),(x_test, y_test) = mnist.load_data()
  x_train, x_test = x_train / 255.0, x_test / 255.0
  
  model = tf.keras.models.Sequential([
    tf.keras.layers.Flatten(input_shape=(28, 28)),
    tf.keras.layers.Dense(128, activation='relu'),
    tf.keras.layers.Dropout(0.2),
    tf.keras.layers.Dense(10, activation='softmax')
  ])
  
  model.compile(optimizer='adam',
    loss='sparse_categorical_crossentropy',
    metrics=['accuracy'])
  
  model.fit(x_train, y_train, epochs=5)
  model.evaluate(x_test, y_test)
  

  Copy

• Run the file in your terminal:

  python3 mnist_example.py
  

  Copy

• Expected output(actual numbers may vary slightly):

  Epoch 1/5
  1875/1875 ━━━━━━━━━━━━━━━━━━━━ 6s 3ms/step - accuracy: 0.8580 - loss: 0.4847     
  Epoch 2/5
  1875/1875 ━━━━━━━━━━━━━━━━━━━━ 5s 3ms/step - accuracy: 0.9551 - loss: 0.1513  
  Epoch 3/5
  1875/1875 ━━━━━━━━━━━━━━━━━━━━ 5s 3ms/step - accuracy: 0.9685 - loss: 0.1068  
  Epoch 4/5
  1875/1875 ━━━━━━━━━━━━━━━━━━━━ 5s 3ms/step - accuracy: 0.9727 - loss: 0.0872  
  Epoch 5/5
  1875/1875 ━━━━━━━━━━━━━━━━━━━━ 5s 3ms/step - accuracy: 0.9773 - loss: 0.0716  
  313/313 ━━━━━━━━━━━━━━━━━━━━ 1s 2ms/step - accuracy: 0.9718 - loss: 0.0870   
  

  Copy

⁠Question and answering

If you have any questions or want to use some special features, please submit an issue or a pull request on openeuler-docker-images⁠.