blob: 6e071ea081b867e2432a384044c47d5c877942ba [file] [log] [blame]
/**
* Licensed to the Apache Software Foundation (ASF) under one
* or more contributor license agreements. See the NOTICE file
* distributed with this work for additional information
* regarding copyright ownership. The ASF licenses this file
* to you 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 "singa/model/layer.h"
#include "./concat.h"
namespace singa {
RegisterLayerClass(singa_concat, Concat);
RegisterLayerClass(singacpp_concat, Concat);
RegisterLayerClass(singacuda_concat, Concat);
RegisterLayerClass(singacl_concat, Concat);
void Concat::Setup(const vector<Shape>& in_shapes, const LayerConf& conf) {
Layer::Setup(in_shapes, conf);
out_sample_shape_.clear();
slice_point_.clear();
axis_ = conf.concat_conf().axis();
CHECK_GE(axis_, 0);
if (axis_ == 0) {
out_sample_shape_ = in_shapes[0];
size_t fea_size = Product(in_shapes[0]);
for (auto& s: in_shapes) {
CHECK_EQ(Product(s), fea_size) << "Feature length of all source samples "
<< "must be the same";
}
} else {
out_sample_shape_ = in_shapes[0];
size_t fea_size = Product(in_shapes[0]) / in_shapes[0][axis_ - 1];
size_t l = 0;
for (auto& s: in_shapes) {
CHECK_GE(s.size(), axis_);
l += s[axis_ - 1];
CHECK_EQ(fea_size, Product(s) / s[axis_ - 1])
<< "Feature length for all axis except axis_ must be the same";
}
out_sample_shape_[axis_ - 1] = l;
}
}
const vector<Tensor> Concat::Forward(int flag, const vector<Tensor>& inputs) {
// TODO(wangwei) check the inputs shape to be the same for all iterations
vector<Tensor> outputs;
slice_point_.clear();
size_t offset = 0;
for (auto& x : inputs) {
offset += x.shape(axis_);
slice_point_.push_back(offset);
}
if (inputs.size() == 1u) {
outputs = inputs;
} else {
outputs.push_back(ConcatOn(inputs, axis_));
}
return outputs;
}
const std::pair<vector<Tensor>, vector<Tensor>> Concat::Backward(
int flag, const vector<Tensor>& grads) {
vector<Tensor> input_grad, param_grad;
CHECK_EQ(grads.size(), 1u) << "Concat layer only have one output tensor.";
size_t last_offset = 0u;
for (auto p : slice_point_) {
input_grad.push_back(SliceOn(grads.at(0), last_offset, p, axis_));
last_offset = p;
}
return std::make_pair(input_grad, param_grad);
}
} // namespace singa