Merge pull request #470 from speechmatics/bugfix-doubleprecision

Fix compilation issues when using KALDI_DOUBLEPRECISION=1
kaldi-asr · Jan 28, 2016 · 75a8334 · 75a8334
2 parents e418aa5 + f587473
commit 75a8334
Show file tree

Hide file tree

Showing 6 changed files with 21 additions and 21 deletions.
diff --git a/src/feat/signal.cc b/src/feat/signal.cc
@@ -34,7 +34,7 @@ void ElementwiseProductOfFft(const Vector<BaseFloat> &a, Vector<BaseFloat> *b) {
 void ConvolveSignals(const Vector<BaseFloat> &filter, Vector<BaseFloat> *signal) {
   int32 signal_length = signal->Dim();
   int32 filter_length = filter.Dim();
-  Vector<float> signal_padded(signal_length + filter_length - 1);
+  Vector<BaseFloat> signal_padded(signal_length + filter_length - 1);
   signal_padded.SetZero();
   for (int32 i = 0; i < signal_length; i++) {
     for (int32 j = 0; j < filter_length; j++) {
@@ -54,11 +54,11 @@ void FFTbasedConvolveSignals(const Vector<BaseFloat> &filter, Vector<BaseFloat>
 
   SplitRadixRealFft<BaseFloat> srfft(fft_length);
 
-  Vector<float> filter_padded(fft_length);
+  Vector<BaseFloat> filter_padded(fft_length);
   filter_padded.Range(0, filter_length).CopyFromVec(filter);
   srfft.Compute(filter_padded.Data(), true);
 
-  Vector<float> signal_padded(fft_length);
+  Vector<BaseFloat> signal_padded(fft_length);
   signal_padded.Range(0, signal_length).CopyFromVec(*signal);
   srfft.Compute(signal_padded.Data(), true);
 
@@ -83,13 +83,13 @@ void FFTbasedBlockConvolveSignals(const Vector<BaseFloat> &filter, Vector<BaseFl
   KALDI_VLOG(1) << "Block size is " << block_length;
   SplitRadixRealFft<BaseFloat> srfft(fft_length);
 
-  Vector<float> filter_padded(fft_length);
+  Vector<BaseFloat> filter_padded(fft_length);
   filter_padded.Range(0, filter_length).CopyFromVec(filter);
   srfft.Compute(filter_padded.Data(), true);
 
-  Vector<float> temp_pad(filter_length - 1);
+  Vector<BaseFloat> temp_pad(filter_length - 1);
   temp_pad.SetZero();
-  Vector<float> signal_block_padded(fft_length);
+  Vector<BaseFloat> signal_block_padded(fft_length);
 
   for (int32 po = 0; po < signal_length; po += block_length) {
     // get a block of the signal

diff --git a/src/lm/kaldi-rnnlm.cc b/src/lm/kaldi-rnnlm.cc
@@ -58,8 +58,8 @@ KaldiRnnlmWrapper::KaldiRnnlmWrapper(
 
 BaseFloat KaldiRnnlmWrapper::GetLogProb(
     int32 word, const std::vector<int32> &wseq,
-    const std::vector<BaseFloat> &context_in,
-    std::vector<BaseFloat> *context_out) {
+    const std::vector<float> &context_in,
+    std::vector<float> *context_out) {
 
   std::vector<std::string> wseq_symbols(wseq.size());
   for (int32 i = 0; i < wseq_symbols.size(); ++i) {
@@ -79,7 +79,7 @@ RnnlmDeterministicFst::RnnlmDeterministicFst(int32 max_ngram_order,
 
   // Uses empty history for <s>.
   std::vector<Label> bos;
-  std::vector<BaseFloat> bos_context(rnnlm->GetHiddenLayerSize(), 1.0f);
+  std::vector<float> bos_context(rnnlm->GetHiddenLayerSize(), 1.0);
   state_to_wseq_.push_back(bos);
   state_to_context_.push_back(bos_context);
   wseq_to_state_[bos] = 0;
@@ -101,7 +101,7 @@ bool RnnlmDeterministicFst::GetArc(StateId s, Label ilabel, fst::StdArc *oarc) {
   KALDI_ASSERT(static_cast<size_t>(s) < state_to_wseq_.size());
 
   std::vector<Label> wseq = state_to_wseq_[s];
-  std::vector<BaseFloat> new_context(rnnlm_->GetHiddenLayerSize());
+  std::vector<float> new_context(rnnlm_->GetHiddenLayerSize());
   BaseFloat logprob = rnnlm_->GetLogProb(ilabel, wseq,
                                          state_to_context_[s], &new_context);
 

diff --git a/src/lm/kaldi-rnnlm.h b/src/lm/kaldi-rnnlm.h
@@ -56,8 +56,8 @@ class KaldiRnnlmWrapper {
   int32 GetEos() const { return eos_; }
 
   BaseFloat GetLogProb(int32 word, const std::vector<int32> &wseq,
-                       const std::vector<BaseFloat> &context_in,
-                       std::vector<BaseFloat> *context_out);
+                       const std::vector<float> &context_in,
+                       std::vector<float> *context_out);
 
  private:
   rnnlm::CRnnLM rnnlm_;
@@ -96,7 +96,7 @@ class RnnlmDeterministicFst
 
   KaldiRnnlmWrapper *rnnlm_;
   int32 max_ngram_order_;
-  std::vector<std::vector<BaseFloat> > state_to_context_;
+  std::vector<std::vector<float> > state_to_context_;
 };
 
 }  // namespace kaldi

diff --git a/src/nnet2/nnet-component.cc b/src/nnet2/nnet-component.cc
@@ -3933,8 +3933,8 @@ void Convolutional1dComponent::Propagate(const ChunkInfo &in_info,
   }
 
   // apply all filters
-  AddMatMatBatched(1.0f, tgt_batch, patch_batch, kNoTrans, filter_params_batch,
-		  kTrans, 1.0f);
+  AddMatMatBatched<BaseFloat>(1.0, tgt_batch, patch_batch, kNoTrans, filter_params_batch,
+		  kTrans, 1.0);
 
   // release memory
   delete filter_params_elem;
@@ -4060,8 +4060,8 @@ void Convolutional1dComponent::Backprop(const ChunkInfo &in_info,
 				    p * num_filters, num_filters)));
     filter_params_batch.push_back(filter_params_elem);  
   }
-  AddMatMatBatched(1.0f, patch_deriv_batch, out_deriv_batch, kNoTrans, 
-		  filter_params_batch, kNoTrans, 0.0f);
+  AddMatMatBatched<BaseFloat>(1.0, patch_deriv_batch, out_deriv_batch, kNoTrans, 
+		  filter_params_batch, kNoTrans, 0.0);
 
   // release memory
   delete filter_params_elem;
@@ -4275,8 +4275,8 @@ void Convolutional1dComponent::Update(const CuMatrixBase<BaseFloat> &in_value,
 				    p * filter_dim, filter_dim)));
   }
 
-  AddMatMatBatched(1.0f, filters_grad_batch, diff_patch_batch, kTrans, patch_batch,
-		  kNoTrans, 1.0f);
+  AddMatMatBatched<BaseFloat>(1.0, filters_grad_batch, diff_patch_batch, kTrans, patch_batch,
+		  kNoTrans, 1.0);
 
   // add the row blocks together to filters_grad
   filters_grad.AddMatBlocks(1.0, filters_grad_blocks_batch);

diff --git a/src/nnet2/nnet-precondition-online-test.cc b/src/nnet2/nnet-precondition-online-test.cc
@@ -307,7 +307,7 @@ void UnitTestPreconditionDirectionsOnline() {
     AssertEqual(trace1, trace2 * gamma2 * gamma2, 1.0e-02);
 
     AssertEqual(Mcopy1, Mcopy2);
-    AssertEqual(row_prod1, row_prod2, 1.0e-02f);
+    AssertEqual<BaseFloat>(row_prod1, row_prod2, 1.0e-02);
     AssertEqual(gamma1, gamma2, 1.0e-02);
 
     // make sure positive definite

diff --git a/src/nnet3/natural-gradient-online-test.cc b/src/nnet3/natural-gradient-online-test.cc
@@ -307,7 +307,7 @@ void UnitTestPreconditionDirectionsOnline() {
     AssertEqual(trace1, trace2 * gamma2 * gamma2, 1.0e-02);
 
     AssertEqual(Mcopy1, Mcopy2);
-    AssertEqual(row_prod1, row_prod2, 1.0e-02f);
+    AssertEqual<BaseFloat>(row_prod1, row_prod2, 1.0e-02);
     AssertEqual(gamma1, gamma2, 1.0e-02);
 
     // make sure positive definite