Machine learning-----> Google Cloud machine learning platform

1. Google Cloud Machine learning Platform Introduction:

The three elements of machine learning are data sources, computing resources, and models. Google has a strong support in these three areas: Google not only has a rich variety of data resources, but also has a strong computer group to provide data storage in the data computing capacity, at the same time, research and implementation of TensorFlow this machine learning, deep learning algorithm library. Based on these backgrounds, Google has also trained a number of practical models that can be applied to commercial software, and developers can directly invoke the appropriate APIs to develop their own business software.

Google Cloud Machine learning is a management platform that aggregates all of these resources (including Google's data resources, computing resources, TensorFlow deep learning algorithm framework, Google's trained models) onto a single platform, This is a highly integrated platform, and it is very convenient to use this platform for machine learning and deep learning.

• The
  • Cloud machine learning handles data in a variety of formats, and cloud machine learning (CML) can access other Google storage, query, and data processing products as plug-ins. Get the data set needed to train the developers to build the model and apply it to the developer's model training process. The data source is Google Cloud Dataproc, a powerful database owned by Google, a global predictive platform that can support tens of thousands of users and massive terabytes of data, enabling the developer-trained model to Plug and Play, which is the Cloud Machie One of the biggest features of the learning platform. The predictive platform integrates Google Cloud Analytics System Cloud Dataflow, allowing developers to access data on Google Cloud storage and bigquery.
  • Cloud machine learning is linked to Google's various cloud services, so developers can easily build their own predictive analytics models using their own training data and train their models on the platform. Cloud machine learning is a cloud-based learning framework that can be used to build and train a custom model of intelligent applications, and currently only has the alpha version. One of the highlights of the
  • Cloud machine Learning management platform, combined with TensorFlow, is the ability to support distributed computing for heterogeneous devices, which can run models automatically on every platform, from phones, single cpu/gpu to distributed systems with hundreds of GPU cards. Developers do not have to spend time on the processing cluster and focus more on model creation. The
  • Cloud Machine learning Platform publishes a number of computer learning models that have been trained by Google, and developers can directly invoke relevant APIs to develop their own projects. This allows you to accelerate the development of a variety of business applications by using the corresponding features provided by Google's machine learning models in their own project. The trained models released on the platform are:
    • "Translate API" (translation)
    • "Cloud Vision AP" (image recognition)
    • "cloud Speech API" (voice Not). Speech recognition API, which uses a neural network model to transform audio into text, supports more than 80 languages, and supports noisy environments and faster intonation. Current use is free and will be charged for the future

2. References

• • How do I use the Google machine learning API?
  • How do I use the TensorFlow library?

　　

Machine learning-----> Google Cloud machine learning platform