nbest for ensemble

1 files changed, 9 insertions, 3 deletions

diff --git a/src/amun/gpu/mblas/nth_element.cu b/src/amun/gpu/mblas/nth_element.cu
index 177ccfb8..c4a426a1 100644
--- a/src/amun/gpu/mblas/nth_element.cu
+++ b/src/amun/gpu/mblas/nth_element.cu
@@ -154,15 +154,21 @@ void NthElement::GetPairs(uint number,
 void NthElement::getValueByKey(std::vector<float>& out, const mblas::Matrix &d_in) const
 {
   // need a model with multiple scorers to test this method
-  assert(false);
+  out.resize(d_breakdown.size());
 
   mblas::VectorWrapper<float> breakdownWrap(d_breakdown);
   const mblas::MatrixWrapper<float> inWrap(d_in);
 
   //gGetValueByKey<<<1, lastN_, 0, stream_>>>
   //  (breakdownWrap, inWrap, h_res_idx, lastN_);
-
-  HANDLE_ERROR( cudaMemcpyAsync(out.data(), d_breakdown.data(), h_res.size() * sizeof(float),
+  /*
+  cerr << "out="
+      << out.size() << " "
+      << d_breakdown.size() << " "
+      << h_res.size()
+      << endl;
+  */
+  HANDLE_ERROR( cudaMemcpyAsync(out.data(), d_breakdown.data(), d_breakdown.size() * sizeof(float),
                                 cudaMemcpyDeviceToHost, mblas::CudaStreamHandler::GetStream()) );
   HANDLE_ERROR( cudaStreamSynchronize(mblas::CudaStreamHandler::GetStream()));
 }