nnet-copy-egs-discriminative.cc File Reference

#include "base/kaldi-common.h"
#include "util/common-utils.h"
#include "hmm/transition-model.h"
#include "nnet2/nnet-example-functions.h"

Include dependency graph for nnet-copy-egs-discriminative.cc:

Go to the source code of this file.

Namespaces
	kaldi
	This code computes Goodness of Pronunciation (GOP) and extracts phone-level pronunciation feature for mispronunciations detection tasks, the reference:
	kaldi::nnet2

Functions
int32	GetCount (double expected_count)
void	AverageConstPart (int32 const_feat_dim, DiscriminativeNnetExample *eg)
int	main (int argc, char *argv[])

Function Documentation

main()

int main	(	int	argc,
		char *	argv[]
	)

Definition at line 65 of file nnet-copy-egs-discriminative.cc.

References kaldi::nnet2::AverageConstPart(), count, SequentialTableReader< Holder >::Done(), ParseOptions::GetArg(), kaldi::nnet2::GetCount(), rnnlm::i, KALDI_LOG, SequentialTableReader< Holder >::Next(), ParseOptions::NumArgs(), ParseOptions::PrintUsage(), kaldi::Rand(), ParseOptions::Read(), ParseOptions::Register(), and SequentialTableReader< Holder >::Value().

                                 {
  try {
    using namespace kaldi;
    using namespace kaldi::nnet2;
    typedef kaldi::int32 int32;
    typedef kaldi::int64 int64;

    const char *usage =
        "Copy examples for discriminative neural\n"
        "network training.  Supports multiple wspecifiers, in\n"
        "which case it will write the examples round-robin to the outputs.\n"
        "\n"
        "Usage:  nnet-copy-egs-discriminative [options] <egs-rspecifier> <egs-wspecifier1> [<egs-wspecifier2> ...]\n"
        "\n"
        "e.g.\n"
        "nnet-copy-egs-discriminative ark:train.degs ark,t:text.degs\n"
        "or:\n"
        "nnet-copy-egs-discriminative ark:train.degs ark:1.degs ark:2.degs\n";
        
    bool random = false;
    int32 srand_seed = 0;
    BaseFloat keep_proportion = 1.0;
    int32 const_feat_dim = 0;

    ParseOptions po(usage);
    po.Register("random", &random, "If true, will write frames to output "
                "archives randomly, not round-robin.");
    po.Register("keep-proportion", &keep_proportion, "If <1.0, this program will "
                "randomly keep this proportion of the input samples.  If >1.0, it will "
                "in expectation copy a sample this many times.  It will copy it a number "
                "of times equal to floor(keep-proportion) or ceil(keep-proportion).");
    po.Register("srand", &srand_seed, "Seed for random number generator "
                "(only relevant if --random=true or --keep-proportion != 1.0)");
    po.Register("const-feat-dim", &const_feat_dim,
                "Dimension of part of features (last dims) which varies little "
                "or not at all with time, and which should be stored as a single "
                "vector for each example rather than in the feature matrix."
                "Useful in systems that use iVectors.  Helpful to save space.");
    
    po.Read(argc, argv);

    srand(srand_seed);
    
    if (po.NumArgs() < 2) {
      po.PrintUsage();
      exit(1);
    }

    std::string examples_rspecifier = po.GetArg(1);

    SequentialDiscriminativeNnetExampleReader example_reader(
        examples_rspecifier);

    int32 num_outputs = po.NumArgs() - 1;
    std::vector<DiscriminativeNnetExampleWriter*> example_writers(num_outputs);
    for (int32 i = 0; i < num_outputs; i++)
      example_writers[i] = new DiscriminativeNnetExampleWriter(
          po.GetArg(i+2));

    
    int64 num_read = 0, num_written = 0, num_frames_written = 0;
    for (; !example_reader.Done(); example_reader.Next(), num_read++) {
      int32 count = GetCount(keep_proportion);
      for (int32 c = 0; c < count; c++) {
        int32 index = (random ? Rand() : num_written) % num_outputs;
        std::ostringstream ostr;
        ostr << num_written;
        if (const_feat_dim == 0) {
          example_writers[index]->Write(ostr.str(),
                                        example_reader.Value());
        } else {
          DiscriminativeNnetExample eg = example_reader.Value();
          AverageConstPart(const_feat_dim, &eg);
          example_writers[index]->Write(ostr.str(), eg);
        }
        num_written++;
        num_frames_written +=
            static_cast<int64>(example_reader.Value().num_ali.size());
      }
    }
    
    for (int32 i = 0; i < num_outputs; i++)
      delete example_writers[i];
    KALDI_LOG << "Read " << num_read << " discriminative neural-network training"
              << " examples, wrote " << num_written << ", consisting of "
              << num_frames_written << " frames.";
    return (num_written == 0 ? 1 : 0);
  } catch(const std::exception &e) {
    std::cerr << e.what() << '\n';
    return -1;
  }
}