ContextQMD
Back to Annotated Deep Learning Paper Implementations

PonderNet: Learning to Ponder

labml_nn/adaptive_computation/ponder_net/readme.md

latest426 B
Original Source

PonderNet: Learning to Ponder

This is a PyTorch implementation of the paper PonderNet: Learning to Ponder.

PonderNet adapts the computation based on the input. It changes the number of steps to take on a recurrent network based on the input. PonderNet learns this with end-to-end gradient descent.