dmlc/mshadow

Wrong device declaration in tensor.h

Sean-Git opened this issue · 0 comments

In file "mshadow/mshadow/tensor.h:840", for cpu device, I think the second param of function AddTakeGradLargeBatch() should be "const Tensor<cpu, 1, IndexType>& sorted" rather than "const Tensor<gpu, 1, IndexType>& sorted".