apaam/mlpack
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Fix for failing tests

Browse Source
This commit is contained in:
Anush V Kini
2020-12-03 19:14:17 +05:30
parent 96b04c2a94
commit a79cf8dd2d
2 changed files with 38 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/mlpack/methods/ann/layer/fast_lstm_impl.hpp
+38 -2
View File
@@ -50,7 +50,16 @@ FastLSTM<InputDataType, OutputDataType>::FastLSTM(const FastLSTM& layer) :
inSize(layer.inSize),
outSize(layer.outSize),
rho(layer.rho),
weights(layer.weights)
forwardStep(layer.forwardStep),
backwardStep(layer.backwardStep),
gradientStep(layer.gradientStep),
weights(layer.weights),
batchSize(layer.batchSize),
batchStep(layer.batchStep),
gradientStepIdx(layer.gradientStepIdx),
grad(layer.grad),
rhoSize(layer.rho),
bpttSteps(layer.bpttSteps)
{
// Nothing to do here.
}
@@ -60,7 +69,16 @@ FastLSTM<InputDataType, OutputDataType>::FastLSTM(FastLSTM&& layer) :
inSize(std::move(layer.inSize)),
outSize(std::move(layer.outSize)),
rho(std::move(layer.rho)),
weights(std::move(layer.weights))
forwardStep(std::move(layer.forwardStep)),
backwardStep(std::move(layer.backwardStep)),
gradientStep(std::move(layer.gradientStep)),
weights(std::move(layer.weights)),
batchSize(std::move(layer.batchSize)),
batchStep(std::move(layer.batchStep)),
gradientStepIdx(std::move(layer.gradientStepIdx)),
grad(std::move(layer.grad)),
rhoSize(std::move(layer.rho)),
bpttSteps(std::move(layer.bpttSteps))
{
// Nothing to do here.
}
@@ -74,7 +92,16 @@ FastLSTM<InputDataType, OutputDataType>::operator=(const FastLSTM& layer)
inSize = layer.inSize;
outSize = layer.outSize;
rho = layer.rho;
forwardStep = layer.forwardStep;
backwardStep = layer.backwardStep;
gradientStep = layer.gradientStep;
weights = layer.weights;
batchSize = layer.batchSize;
batchStep = layer.batchStep;
gradientStepIdx = layer.gradientStepIdx;
grad = layer.grad;
rhoSize = layer.rho;
bpttSteps = layer.bpttSteps;
}
return *this;
}
@@ -88,7 +115,16 @@ FastLSTM<InputDataType, OutputDataType>::operator=(FastLSTM&& layer)
inSize = std::move(layer.inSize);
outSize = std::move(layer.outSize);
rho = std::move(layer.rho);
forwardStep = std::move(layer.forwardStep);
backwardStep = std::move(layer.backwardStep);
gradientStep = std::move(layer.gradientStep);
weights = std::move(layer.weights);
batchSize = std::move(layer.batchSize);
batchStep = std::move(layer.batchStep);
gradientStepIdx = std::move(layer.gradientStepIdx);
grad = std::move(layer.grad);
rhoSize = std::move(layer.rho);
bpttSteps = std::move(layer.bpttSteps);
}
return *this;
}
src/mlpack/methods/ann/rnn_impl.hpp
-5
View File
@@ -71,11 +71,6 @@ RNN<OutputLayerType, InitializationRuleType, CustomLayers...>::RNN(
network.network[i]));
boost::apply_visitor(resetVisitor, this->network.back());
}
ResetCells();
if (parameter.is_empty())
{
ResetParameters();
}
}
template<typename OutputLayerType, typename InitializationRuleType,