ginn-org/ginn

Issues

• Ensure speed parity between mnist-layers and mnist examples

  #44 opened a year ago by oir
  0

• Investigate meson

  #36 opened 2 years ago by oir
  1

• Investigate incorporating FasterTransformer

  #39 opened 2 years ago by oir
  0

• Deprecate core.h

  #35 opened 2 years ago by oir
  0

• Gpu-enabled cpu runs are slow

  #26 opened 2 years ago by oir
  1

• Determine semantics of "grad"/"backward" for integral scalar types

  #29 opened 2 years ago by oir
  1

• Separate Backward Graph

  #19 opened 2 years ago by oir
  1

• Consider redesigning operator=() for tensors

  #28 opened 2 years ago by oir
  1

• Investigate removing Node::set_ins()

  #25 opened 2 years ago by oir
  0

• Check devices of inputs when calling forward

  #24 opened 2 years ago by oir
  0

• Investigate using cuda streams for async GPU operations

  #23 opened 2 years ago by oir
  0

• Visualize computation graph with dotviz

  #22 opened 2 years ago by oir
  0

• AddScalar and ProdScalar with scalar output nodes

  #21 opened 2 years ago by oir
  0

• Run clang-tidy

  #20 opened 2 years ago by oir
  0

• Composing expressions (of unary functions in particular)

  #18 opened 2 years ago by oir
  0

• Investigate possibility of an eager mode

  #17 opened 2 years ago by oir
  0

• Add operator<< support for Tensors, Shapes, .. ?

  #16 opened 2 years ago by oir
  0

• Add detailed error messages (ex: dimension mismatch)

  #15 opened 2 years ago by oir
  0

• Check feature parity between Python and C++

  #14 opened 2 years ago by oir
  0

• Add more update rules

  #13 opened 2 years ago by oir
  0

• Add language modeling example with beam search

  #12 opened 2 years ago by oir
  0

• Add Crf

  #11 opened 2 years ago by oir
  0

• Add API documentation

  #10 opened 2 years ago by oir
  0

• Add autobatching for Conv nodes

  #9 opened 2 years ago by oir
  0

• Investigate Onnx support

  #8 opened 2 years ago by oir
  0

• Investigate Cudnn support

  #7 opened 2 years ago by oir
  0

• Add modules for automagically downloading datasets

  #6 opened 2 years ago by oir
  0

• Make headers self includeable

  #5 opened 2 years ago by oir
  0

• Add Elman RNNs and GRUs

  #4 opened 2 years ago by oir
  0

• Investigate single header only option

  #3 opened 2 years ago by oir
  0

• Investigate if there is any speed benefits of unrolling Eigen summation in Affine node

  #2 opened 2 years ago by oir
  0

• Investigate using parent c instead of children's c when computing o in TreeLstm

  #1 opened 2 years ago by oir
  0