• Tensor subclasses and PT2

  • Feb 24 2024
  • Duración: 13 m
  • Podcast

  • Resumen

  • Tensor subclasses allow you to add extend PyTorch with new types of tensors without having to write any C++. They have been used to implement DTensor, FP8, Nested Jagged Tensor and Complex Tensor. Recent work by Brian Hirsh means that we can compile tensor subclasses in PT2, eliminating their overhead. The basic mechanism by which this compilation works is a desugaring process in AOTAutograd. There are some complications involving views, dynamic shapes and tangent metadata mismatch.
    Más Menos
activate_primeday_promo_in_buybox_DT

Lo que los oyentes dicen sobre Tensor subclasses and PT2

Calificaciones medias de los clientes

Reseñas - Selecciona las pestañas a continuación para cambiar el origen de las reseñas.