search

fengggli / gpu-computing-materials

A simple deep learning framework that optimizes task scheduling and memory usage on different CPU/GPU architectures.
1 stars 0 forks source link

[pooling layer]: add forward and backward #14

Closed fengggli closed 5 years ago

fengggli commented 5 years ago

added

  1. Tensor operation: tensor_make_empty_with_dim
  2. add pooling layer forward and backward
  3. add partial test