Crate caffe2op_tt

source ·

Structs

  • | TODO: Complete after verifying utility | of TT-layer’s forward pass. |
  • | The TT-layer serves as a low-rank decomposition | of a fully connected layer. | | The inputs are the same as to a fully connected | layer, but the number of parameters | are greatly reduced and forward computation | time can be drastically reduced especially | for layers with large weight matrices. | | The multiplication is computed as a | product of the input vector with each | of the cores that make up the TT layer. | | Given the input sizes (inp_sizes), | output sizes(out_sizes), and the ranks | of each of the cores (tt_ranks), the | ith core will have size: | | inp_sizes[i] * tt_ranks[i] * | tt_ranks[i + 1] * out_sizes[i]. | | The complexity of the computation is | dictated by the sizes of inp_sizes, | out_sizes, and tt_ranks, where there | is the trade off between accuracy of | the low-rank decomposition and the | speed of the computation. |