tencent cloud

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 machine learning and developers easily build and deploy machine learning-powered applications.

• Easy model building
  Build and train ML models easily using intuitive high-level APIs like Keras with eager execution, which makes for immediate model iteration and easy debugging.
• Reliable machine learning production anywhere
  Easily train and deploy models in the cloud, on-prem, in the browser, or on-device no matter what language you use.
• Powerful experimentation for research
  A simple and flexible architecture to take new ideas from concept to code, to state-of-the-art models, and to publication faster.

TensorFlow Architecture

• Client
  It defines the computation process as a data flow graph and initializes graph execution by using _Session_.
• Distributed master
  It prunes specific subgraphs in a graph, i.e, parameters defined in Session.run(), partitions a subgraph into multiple parts that run in different processes and devices, and distributes the graph parts to different worker services, which initialize subgraph computation.
• Worker service (one for each task)
  It schedules the execution of graph operations by using kernel implementations appropriate to the available hardware (CPUs, GPUs, etc.) and sends/receives operation results to/from other worker services.
• Kernel implementation
  It performs the computation for individual graph operations.

EMR's Support for TensorFlow

• TensorFlow version: v1.14.0
• Currently, TensorFlow can only run on CPU models instead of GPU models.
• TensorFlowOnSpark can be used for distributed training.

TensorFlow Development Sample

Write code: test.py

import tensorflow as tf
hello = tf.constant('Hello, TensorFlow!')
sess = tf.Session()
print sess.run(hello)
a = tf.constant(10)
b = tf.constant(111)
print sess.run(a+b)
exit()

Run the following command:

python test.py

For more usage, please visit the TensorFlow official website.