add base class for seqlastin/max/average layer (#187)
luotao1
8 years ago
committed by
emailweixu
parent
76fb74dc2e
commit
e1f57bfd66
@ -0,0 +1,84 @@
|
||||
/* Copyright (c) 2016 Baidu, Inc. All Rights Reserve.
|
||||
|
||||
Licensed under the Apache License, Version 2.0 (the "License");
|
||||
you may not use this file except in compliance with the License.
|
||||
You may obtain a copy of the License at
|
||||
|
||||
http://www.apache.org/licenses/LICENSE-2.0
|
||||
|
||||
Unless required by applicable law or agreed to in writing, software
|
||||
distributed under the License is distributed on an "AS IS" BASIS,
|
||||
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
||||
See the License for the specific language governing permissions and
|
||||
limitations under the License. */
|
||||
|
||||
#include "paddle/utils/Logging.h"
|
||||
#include "SequencePoolLayer.h"
|
||||
|
||||
namespace paddle {
|
||||
|
||||
bool SequencePoolLayer::init(const LayerMap& layerMap,
|
||||
const ParameterMap& parameterMap) {
|
||||
/* Initialize the basic parent class */
|
||||
Layer::init(layerMap, parameterMap);
|
||||
|
||||
// seqlastins/max/average layer should have exactly 1 input
|
||||
CHECK_EQ(1U, inputLayers_.size());
|
||||
|
||||
/* initialize biases_ */
|
||||
if (biasParameter_.get() != NULL) {
|
||||
biases_ = std::unique_ptr<Weight>(new Weight(1, getSize(), biasParameter_));
|
||||
}
|
||||
// transform to which sequence type
|
||||
if (config_.trans_type() == "non-seq") {
|
||||
type_ = kNonSeq;
|
||||
} else if (config_.trans_type() == "seq") {
|
||||
type_ = kSeq;
|
||||
} else {
|
||||
LOG(FATAL) << "Unknown trans_type: " << config_.trans_type();
|
||||
}
|
||||
setNeedSequenceInfo(false);
|
||||
return true;
|
||||
}
|
||||
|
||||
void SequencePoolLayer::forward(PassType passType) {
|
||||
Layer::forward(passType);
|
||||
|
||||
const Argument& input = getInput(0);
|
||||
newBatchSize_ = type_ ? input.getNumSubSequences() : input.getNumSequences();
|
||||
size_t dim = getSize();
|
||||
// check
|
||||
CHECK_EQ(dim, input.value->getWidth());
|
||||
startPositions_ =
|
||||
type_ ? input.subSequenceStartPositions : input.sequenceStartPositions;
|
||||
auto starts = startPositions_->getVector(false);
|
||||
CHECK_EQ(starts->getData()[newBatchSize_], input.getBatchSize());
|
||||
CHECK_EQ(newBatchSize_, starts->getSize() - 1);
|
||||
|
||||
resetOutput(newBatchSize_, dim);
|
||||
if (type_) {
|
||||
CHECK(input.subSequenceStartPositions)
|
||||
<< "when trans_type = seq, input must hasSubseq";
|
||||
}
|
||||
/* If type_ = kNonSeq, both seq has or not has sub-seq degrade to a non-seq,
|
||||
* thus, in this case, output_ has no sequenceStartPositions.
|
||||
* If type_ = kSeq, seq has sub-seq degrades to a seq, thus, only in this
|
||||
* case, we should compute the new sequenceStartPositions.
|
||||
*/
|
||||
if (type_) {
|
||||
output_.degradeSequence(input, useGpu_);
|
||||
}
|
||||
}
|
||||
|
||||
void SequencePoolLayer::backward(const UpdateCallback& callback) {
|
||||
/* Do derivation */ { backwardActivation(); }
|
||||
|
||||
if (biases_ && biases_->getWGrad()) {
|
||||
biases_->getWGrad()->collectBias(*getOutputGrad(), 1);
|
||||
|
||||
// Increasing the number of gradient
|
||||
biases_->getParameterPtr()->incUpdate(callback);
|
||||
}
|
||||
}
|
||||
|
||||
} // namespace paddle
|
||||
@ -0,0 +1,57 @@
|
||||
/* Copyright (c) 2016 Baidu, Inc. All Rights Reserve.
|
||||
|
||||
Licensed under the Apache License, Version 2.0 (the "License");
|
||||
you may not use this file except in compliance with the License.
|
||||
You may obtain a copy of the License at
|
||||
|
||||
http://www.apache.org/licenses/LICENSE-2.0
|
||||
|
||||
Unless required by applicable law or agreed to in writing, software
|
||||
distributed under the License is distributed on an "AS IS" BASIS,
|
||||
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
||||
See the License for the specific language governing permissions and
|
||||
limitations under the License. */
|
||||
|
||||
#pragma once
|
||||
|
||||
#include "Layer.h"
|
||||
#include "paddle/math/Matrix.h"
|
||||
|
||||
namespace paddle {
|
||||
/**
|
||||
* A base layer for SequenceLastInstanceLayer/AverageLayer/MaxLayer.
|
||||
*
|
||||
* Input: one or more sequences. Each sequence contains some instances.
|
||||
* If SequenceLevel = kNonSeq:
|
||||
* Output: output size is the number of input sequences (NOT input instances)
|
||||
* output[i] = seqlastin/average/max_{for each instance in this
|
||||
* sequence}{input[i]}
|
||||
* If SequenceLevel = kSeq:
|
||||
* Check input sequence must has sub-sequence
|
||||
* Output: output size is the number of input sub-sequences
|
||||
* output[i] = seqlastin/average/max_{for each instance in this
|
||||
* sub-sequence}{input[i]}
|
||||
*
|
||||
* The config file api is pooling_layer.
|
||||
*/
|
||||
|
||||
class SequencePoolLayer : public Layer {
|
||||
protected:
|
||||
int type_;
|
||||
std::unique_ptr<Weight> biases_;
|
||||
enum SequenceLevel { kNonSeq = 0, kSeq = 1 };
|
||||
size_t newBatchSize_;
|
||||
ICpuGpuVectorPtr startPositions_;
|
||||
|
||||
public:
|
||||
explicit SequencePoolLayer(const LayerConfig& config) : Layer(config) {}
|
||||
|
||||
virtual ~SequencePoolLayer() {}
|
||||
|
||||
bool init(const LayerMap& layerMap, const ParameterMap& parameterMap);
|
||||
|
||||
void forward(PassType passType);
|
||||
void backward(const UpdateCallback& callback = nullptr);
|
||||
};
|
||||
|
||||
} // namespace paddle
|
||||
Loading…
Reference in new issue