Home > Categories > Machine learning and neural networks > Model Deployment > TensorFlow Serving

TensorFlow Serving

Rating:

4.7 / 5.0

Tags

Machine Learning, Model Deployment, Serving, TensorFlow, Open Source, MLOps, Prediction API

Categories:
Machine learning and neural networks
Creator Google
Date 2016-03-11
Platforms Linux, Docker, Kubernetes, Cloud
Status Active
Website tensorflow.org
Price Model Free (library)
Sections:
DL FrameworksModel Deployment

Pricing Details

Free (open source). Costs may apply when using compute resources for hosting.

Features

High-performance model serving, supports TensorFlow models, REST and gRPC APIs, model versioning, scalability, open-source.

Integrations

Integration with TensorFlow models. API for integration into any applications. Can be deployed in Docker, Kubernetes, Google Cloud, AWS, Azure.

Preview

TensorFlow Serving is a flexible, high-performance serving system for machine learning models in production. It is developed by Google and makes it easy to deploy trained TensorFlow models (and other compatible models) to get predictions via APIs (REST and gRPC). TensorFlow Serving supports model versioning, allowing for easy model updates and rollbacks. The system is optimized for high throughput and low latency, making it ideal for applications requiring fast, real-time predictions. It is open-source and can be deployed in various environments, including cloud platforms and on-premises servers.