101.school
CoursesAbout
Search...⌘K
Generate a course with AI...

    Tensorflow

    Receive aemail containing the next unit.
    • Introduction to Tensorflow
      • 1.1Understanding the Basics of Tensorflow
      • 1.2Working with Tensorflow Constants, Variables, and Placeholders
      • 1.3Understanding Tensorflow Sessions
      • 1.4Concepts of Graphs in Tensorflow
    • Deep Learning and Neural Networks
      • 2.1Deep Learning Fundamentals
      • 2.2Introduction to Neural Networks
      • 2.3Building a Neural Network in Tensorflow
      • 2.4Implementing Neural Networks for Regression problems
    • Convolutional Neural Networks (CNN) and Recurrent Neural Networks (RNN)
      • 3.1Introduction to Convolutional Neural Networks
      • 3.2Practical use-cases of CNN
      • 3.3Understanding Recurrent Neural Networks
      • 3.4Practical use-cases of RNN
    • Advanced Topics in Tensorflow
      • 4.1TFRecords and TensorBoard
      • 4.2Saving and Restoring Tensorflow Models
      • 4.3Tensorflow Lite and Tensorflow.js
      • 4.4Tensorflow Extended (TFX)

    Advanced Topics in Tensorflow

    Understanding Tensorflow Extended (TFX)

    machine learning software library

    Machine learning software library.

    Tensorflow Extended (TFX) is an end-to-end platform for deploying production-ready machine learning (ML) pipelines. It provides a configuration framework and shared libraries to integrate common components needed to define, launch, and monitor your ML system.

    Components of Tensorflow Extended

    TFX consists of several components that work together to manage the lifecycle of a machine learning model. These components include:

    • ExampleGen: This is the first component of any TFX pipeline. It ingests and splits the input dataset.
    • StatisticsGen: This component computes statistics for the dataset for data analysis and for use in downstream components.
    • SchemaGen: This auto-generates a schema for the input data based on the statistics from StatisticsGen.
    • ExampleValidator: This component identifies anomalies and missing values in the dataset.
    • Transform: This component performs feature engineering on the dataset.
    • Trainer: This component trains the model using user-provided TensorFlow code.
    • Evaluator: This component performs deep analysis of the training results.
    • ModelValidator: This component ensures that the model is "good enough" to be pushed to production.
    • Pusher: This component deploys the model in a serving infrastructure.

    Data Validation and Transformation in TFX

    Data validation in TFX is performed by the ExampleValidator component, which identifies anomalies and missing values in the dataset. The anomalies are identified based on the expectations defined in the schema and the statistics of the data.

    The Transform component performs feature engineering on the dataset. It uses the schema and the user-provided TensorFlow code to transform the raw data into features suitable for ML model training.

    Training, Tuning, and Validating Models with TFX

    The Trainer component trains the model using user-provided TensorFlow code. It uses the transformed data and schema from the Transform component, and trains the model based on this. The output of this component is a SavedModel.

    The Evaluator component performs deep analysis of the training results. It uses the validation set and the trained model to compute evaluation metrics, which can be visualized in TensorBoard.

    The ModelValidator component ensures that the model is "good enough" to be pushed to production. It compares the new model with the current model in production and checks if the new model is better.

    Deploying Models with TFX

    The Pusher component deploys the model in a serving infrastructure. If the ModelValidator has marked the model as "good enough", the Pusher component deploys the model to a serving infrastructure (like TensorFlow Serving or TensorFlow Lite).

    Continuous Integration and Delivery with TFX

    TFX supports continuous integration and delivery, allowing for regular model training and deployment, ensuring that models are up-to-date and improving over time. This is achieved through TFX's ability to create reproducible pipelines, which can be triggered based on new data, and its robust validation and testing components.

    In conclusion, Tensorflow Extended (TFX) is a powerful tool for managing the lifecycle of a machine learning model, providing components for each step of the process, from data ingestion and validation, to model training and deployment.

    Test me
    Practical exercise
    Further reading

    Hey there, any questions I can help with?

    Sign in to chat