Welcome to the MediaPipe Solutions Preview. We are excited to see what you develop with this new offering and read your feedback.
On-device machine learning is evolving. Use cases are more complex, potentially requiring numerous multimodal models, domain-specific processes, and tightly integrated pipelines for end-to-end performance. This new MediaPipe Solutions is a unification of several existing tools: MediaPipe Solutions, TensorFlow Lite Task Library, and TensorFlow Lite Model Maker.
MediaPipe Tasks: Low-code API to create and deploy advanced ML solutions across platforms.
- What's new: goes beyond single model inference with end-to-end optimized pipeline performance
- What's unchanged: same easy workflow to deploy a solution with a few lines of code
MediaPipe Model Maker: Low-code API to customize solutions using your own data.
- What's new: refactored to support robust use cases and MediaPipe Studio infrastructure
- What's unchanged: same easy workflow to customize a model with a few lines of code
MediaPipe Studio: Visualize and benchmark solutions.
- What's new: new way to upload and visualize supported solutions
- What's unchanged: built with Google ML expertise for performance and scalability
Explore the examples and start building your ML solutions with the developer guides.