ROperator_Split.hxx

Go to the documentation of this file.

#ifndef TMVA_SOFIE_ROPERATOR_Split
#define TMVA_SOFIE_ROPERATOR_Split
 
#include "TMVA/SOFIE_common.hxx"
#include "TMVA/ROperator.hxx"
#include "TMVA/RModel.hxx"
 
#include <sstream>
 
namespace TMVA{
namespace Experimental{
namespace SOFIE{
 
 
class ROperator_Split final : public ROperator
{
 
private:
 
   int fAxis  = 0;
   std::string fNX;
   std::string fNSplit;
   std::vector<std::string> fNYs;
   std::vector<Dim> fInputShape;
   std::vector<int64_t> fSplit;
   std::vector<std::vector<Dim>> fOutputShapes;
 
 
 
public:
   ROperator_Split(){}
   ROperator_Split(const std::string & nameX, const std::string & nameS,  int axis, const std::vector<std::string> &  namesY):
      fAxis(axis), fNX(UTILITY::Clean_name(nameX)), fNSplit(UTILITY::Clean_name(nameS)) {
         fNYs.reserve(namesY.size());
         for (auto & name : namesY)
            fNYs.push_back(UTILITY::Clean_name(name));
 
         fInputTensorNames = { fNX };
         fOutputTensorNames.resize(fNYs.size());
         std::transform(fNYs.begin(), fNYs.end(), fOutputTensorNames.begin(),
                   [](const std::string& s) -> std::string_view { return s; });
      }
 
   void Initialize(RModel& model) override {
      if (model.CheckIfTensorAlreadyExist(fNX) == false){   //input must be a graph input, or already initialized intermediate tensor
         throw std::runtime_error("TMVA SOFIE Split Op Input Tensor is not found in model");
      }
      fInputShape = model.GetDimTensorShape(fNX);
 
      // correct for negative axis
      if (fAxis < 0) fAxis += fInputShape.size();
      if (fAxis < 0 || fAxis >= static_cast<int>(fInputShape.size()) )
         throw std::runtime_error("TMVA SOFIE Split - invalid axis " + std::to_string(fAxis));
 
      // support for time being split in axis whi are defined not parametrics
      if (fInputShape[fAxis].isParam)
         throw std::runtime_error("TMVA SOFIE Split - splitting in dynamic axis is not supported");
 
      size_t origValue = fInputShape[fAxis].dim;
 
      // compute output shapes
      size_t nsplit = fNYs.size();
      // case split tensor is empty
      if (fNSplit.empty()) {
         int64_t splitValue = 0;
         if (origValue % nsplit == 0) {
            splitValue = origValue/nsplit;
            fSplit = std::vector<int64_t>(nsplit, splitValue);
         } else {
            // case of not equal splitting
            splitValue = std::ceil(double(origValue)/nsplit);
            fSplit = std::vector<int64_t>(nsplit-1, splitValue);
            fSplit.push_back(origValue % splitValue);
         }
      } else {
         // NB : in this case we could support dynamic split axes
         // get split tensor values
         if (!model.IsInitializedTensor(fNSplit))
            throw std::runtime_error("TMVA SOFIE Split - non-initialized split tensors are not supported");
         auto splitShape =  model.GetTensorShape(fNSplit);
         if (splitShape.size() != 1 || splitShape[0] != nsplit)
            throw std::runtime_error("TMVA SOFIE Split - split input tensor has invalid shape");
         auto split_data = static_cast<int64_t *>(model.GetInitializedTensorData(fNSplit).get());
         fSplit = std::vector<int64_t>(split_data, split_data + nsplit);
      }
      // compute now the output shapes
      size_t tot_split = 0;
      for (size_t i = 0; i < fNYs.size(); i++) {
         std::vector<Dim> outputShape = fInputShape;
         outputShape[fAxis] = Dim{ static_cast<size_t>(fSplit[i]) };
         tot_split += fSplit[i];
         model.AddIntermediateTensor(fNYs[i], model.GetTensorType(fNX), outputShape);
         fOutputShapes.push_back(outputShape);
      }
      if (tot_split != origValue)
         throw std::runtime_error("TMVA SOFIE Split - Sum of split sizes must match the input dimension along the axis");
 
 
      if (model.Verbose()) {
         std::cout << "Split - input shape " << ConvertDimShapeToString(fInputShape) << " --> ";
         for (auto & s : fOutputShapes)
            std::cout << ConvertDimShapeToString(s) << "  ";
         std::cout << std::endl;
      }
   }
 
 
   std::string Generate(std::string OpName) override {
      OpName = "op_" + OpName;
      if (fOutputShapes.empty()){
         throw std::runtime_error("TMVA SOFIE Operator Split called to Generate without being initialized first");
      }
 
      auto input_strides =  UTILITY::ComputeStrideFromShape(fInputShape);
 
      // generate now the code for split
      std::stringstream out;
      out << "\n" << SP << "//------ Split\n";
      out << SP << "size_t " << OpName << "_axis_offset = 0;\n";
      // unroll the loop on split outputs
      for (size_t i = 0; i < fNYs.size(); i++)  {
         auto length = ConvertDimShapeToLength(fOutputShapes[i]);
         auto output_strides = UTILITY::ComputeStrideFromShape(fOutputShapes[i]);
 
         out << SP << "for (int id = 0; id < " << length << " ; id++){\n";
         // convert output index to input index
         out << SP << SP << "int input_index = 0;\n";
         out << SP << SP << "int remaining = id;\n";
         // loop on dimensions to compute the input indices(unroll this loop)
         for (size_t k = 0; k < fOutputShapes[i].size(); ++k) {
            out << SP << SP << "// dim " << k << "\n";
            if (k < fOutputShapes[i].size()-1) {
               out << SP << SP << "input_index += (int(remaining / " << output_strides[k] << ")";
               // for the split axis we need to consider the offset in the splits when converting to input coordinates
               if (k == static_cast<size_t>(fAxis) && i > 0)
                  out << " + " << OpName << "_axis_offset";
               out << ") * " << input_strides[k] << ";\n";
               out << SP << SP  << "remaining %= " << output_strides[k] << ";\n";
            } else {
               // for last dims all strides are one
               out << SP << SP << "input_index += remaining";
               if (k == static_cast<size_t>(fAxis) && i > 0)
                  out << " + " << OpName << "_axis_offset";
               out << ";\n\n";
            }
         }
 
         out << SP << SP  << "tensor_" << fNYs[i] << "[id] = tensor_" << fNX <<"[input_index];\n";
         out << SP << "}\n";
         if (i < fNYs.size()-1) out << SP << OpName << "_axis_offset += " << fSplit[i] << ";\n";
      }
      return out.str();
   }
 
};
 
}//SOFIE
}//Experimental
}//TMVA
 
 
#endif //TMVA_SOFIE_ROPERATOR_Swish