TensorFlow Lite

Latest news

1. 29 Oct
   Google adds faster GPU inference and smaller app sizes with LiteRT in Play services

   Google upgraded LiteRT in Google Play services to support a faster GPU inference stack and a new CompiledModel API that replaces the legacy Interpreter API. Apps that switch to the shared runtime in Play services can be smaller (about 8 MB in Google testing) and benefit from roughly 1.4x faster GPU inference than the legacy GPU delegate.

   Developers · 1 source

   

Articles

No articles about TensorFlow Lite yet.