Add support for dynamic autodiff
Change-Id: I17d573696172ab691a9653db99a620e4bc1bd0d0
diff --git a/include/ceres/dynamic_autodiff_cost_function.h b/include/ceres/dynamic_autodiff_cost_function.h
new file mode 100644
index 0000000..0bc18eb
--- /dev/null
+++ b/include/ceres/dynamic_autodiff_cost_function.h
@@ -0,0 +1,189 @@
+// Ceres Solver - A fast non-linear least squares minimizer
+// Copyright 2012 Google Inc. All rights reserved.
+// http://code.google.com/p/ceres-solver/
+//
+// Redistribution and use in source and binary forms, with or without
+// modification, are permitted provided that the following conditions are met:
+//
+// * Redistributions of source code must retain the above copyright notice,
+// this list of conditions and the following disclaimer.
+// * Redistributions in binary form must reproduce the above copyright notice,
+// this list of conditions and the following disclaimer in the documentation
+// and/or other materials provided with the distribution.
+// * Neither the name of Google Inc. nor the names of its contributors may be
+// used to endorse or promote products derived from this software without
+// specific prior written permission.
+//
+// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+// AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+// IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+// ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
+// LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+// CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+// SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+// INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+// CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+// ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+// POSSIBILITY OF SUCH DAMAGE.
+//
+// Author: mierle@gmail.com (Keir Mierle)
+// sameeragarwal@google.com (Sameer Agarwal)
+// thadh@gmail.com (Thad Hughes)
+//
+// This autodiff implementation differs from the one found in
+// autodiff_cost_function.h by supporting autodiff on cost functions with
+// variable numbers of parameters with variable sizes. With the other
+// implementation, all the sizes (both the number of parameter blocks and the
+// size of each block) must be fixed at compile time.
+//
+// The functor API differs slightly from the API for fixed size autodiff; the
+// expected interface for the cost functors is:
+//
+// struct MyCostFunctor {
+// template<typename T>
+// bool operator()(const* const* T parameters, T* residuals) const {
+// // Use parameters[i] to access the i'th parameter block.
+// }
+// }
+//
+// Since the sizing of the parameters is done at runtime, you must also specify
+// the sizes after creating the dynamic autodiff cost function. For example:
+//
+// DynamicAutoDiffCostFunction<MyCostFunctor, 3> cost_function(
+// new MyCostFunctor());
+// cost_function.AddParameterBlock(5);
+// cost_function.AddParameterBlock(10);
+// cost_function.SetNumResiduals(21);
+//
+// Under the hood, the implementation evaluates the cost function multiple
+// times, computing a small set of the derivatives (four by default, controlled
+// by the Stride template parameter) with each pass. There is a tradeoff with
+// the size of the passes; you may want to experiment with the stride.
+
+#ifndef CERES_PUBLIC_DYNAMIC_AUTODIFF_COST_FUNCTION_H_
+#define CERES_PUBLIC_DYNAMIC_AUTODIFF_COST_FUNCTION_H_
+
+#include <cmath>
+#include <numeric>
+#include <glog/logging.h>
+#include "ceres/cost_function.h"
+#include "ceres/internal/scoped_ptr.h"
+#include "ceres/jet.h"
+
+namespace ceres {
+
+template <typename CostFunctor, int Stride = 4>
+class DynamicAutoDiffCostFunction : public CostFunction {
+ public:
+ DynamicAutoDiffCostFunction(CostFunctor* functor)
+ : functor_(functor) {}
+
+ virtual ~DynamicAutoDiffCostFunction() {}
+
+ void AddParameterBlock(int size) {
+ mutable_parameter_block_sizes()->push_back(size);
+ }
+
+ void SetNumResiduals(int num_residuals) {
+ set_num_residuals(num_residuals);
+ }
+
+ virtual bool Evaluate(double const* const* parameters,
+ double* residuals,
+ double** jacobians) const {
+ CHECK_GT(num_residuals(), 0)
+ << "You must call DynamicAutoDiffCostFunction::SetNumResiduals() "
+ << "before DynamicAutoDiffCostFunction::Evaluate().";
+
+ if (jacobians == NULL) {
+ return (*functor_)(parameters, residuals);
+ }
+
+ // The difficulty with Jets, as implemented in Ceres, is that they were
+ // originally designed for strictly compile-sized use. At this point, there
+ // is a large body of code that assumes inside a cost functor it is
+ // acceptable to do e.g. T(1.5) and get an appropriately sized jet back.
+ //
+ // Unfortunately, it is impossible to communicate the expected size of a
+ // dynamically sized jet to the static instantiations that existing code
+ // depends on.
+ //
+ // To work around this issue, the solution here is to evaluate the
+ // jacobians in a series of passes, each one computing Stripe *
+ // num_residuals() derivatives. This is done with small, fixed-size jets.
+ const int num_parameter_blocks = parameter_block_sizes().size();
+ const int num_parameters = std::accumulate(parameter_block_sizes().begin(),
+ parameter_block_sizes().end(),
+ 0);
+
+ // Allocate scratch space for the strided evaluation.
+ vector<Jet<double, Stride> > input_jets(num_parameters);
+ vector<Jet<double, Stride> > output_jets(num_residuals());
+
+ // Make the parameter pack that is sent to the functor (reused).
+ vector<Jet<double, Stride>* > jet_parameters(num_parameter_blocks);
+ for (int i = 0, parameter_cursor = 0; i < num_parameter_blocks; ++i) {
+ jet_parameters[i] = &input_jets[parameter_cursor];
+ for (int j = 0; j < parameter_block_sizes()[i]; ++j, parameter_cursor++) {
+ input_jets[parameter_cursor].a = parameters[i][j];
+ }
+ }
+
+ // Evaluate all of the strides. Each stride is a chunk of the derivative to
+ // evaluate, typically some size proportional to the size of the SIMD
+ // registers of the CPU.
+ int num_strides = int(ceil(num_parameters / float(Stride)));
+ for (int pass = 0; pass < num_strides; ++pass) {
+ const int start_derivative_section = pass * Stride;
+ const int end_derivative_section = std::min((pass + 1) * Stride,
+ num_parameters);
+ // Set most of the jet components to zero, except for the active
+ // parameters, which occur in a contiguos block of size Stride.
+ for (int i = 0, parameter_cursor = 0; i < num_parameter_blocks; ++i) {
+ for (int j = 0; j < parameter_block_sizes()[i];
+ ++j, parameter_cursor++) {
+ input_jets[parameter_cursor].v.setZero();
+ if (parameter_cursor >= start_derivative_section &&
+ parameter_cursor < end_derivative_section) {
+ input_jets[parameter_cursor]
+ .v[parameter_cursor - start_derivative_section] = 1.0;
+ }
+ }
+ }
+
+ if (!(*functor_)(&jet_parameters[0], &output_jets[0])) {
+ return false;
+ }
+
+ // Copy the pieces of the jacobians into their final place.
+ for (int i = 0, parameter_cursor = 0; i < num_parameter_blocks; ++i) {
+ for (int j = 0; j < parameter_block_sizes()[i];
+ ++j, parameter_cursor++) {
+ if (parameter_cursor >= start_derivative_section &&
+ parameter_cursor < end_derivative_section) {
+ for (int k = 0; k < num_residuals(); ++k) {
+ jacobians[i][k * parameter_block_sizes()[i] + j] =
+ output_jets[k].v[parameter_cursor - start_derivative_section];
+ }
+ }
+ }
+ }
+
+ // Only copy the residuals over once (even though we compute them on
+ // every loop).
+ if (pass == num_strides - 1) {
+ for (int k = 0; k < num_residuals(); ++k) {
+ residuals[k] = output_jets[k].a;
+ }
+ }
+ }
+ return true;
+ }
+
+ private:
+ internal::scoped_ptr<CostFunctor> functor_;
+};
+
+} // namespace ceres
+
+#endif // CERES_PUBLIC_DYNAMIC_AUTODIFF_COST_FUNCTION_H_
diff --git a/internal/ceres/CMakeLists.txt b/internal/ceres/CMakeLists.txt
index f9329e9..b30f0cc 100644
--- a/internal/ceres/CMakeLists.txt
+++ b/internal/ceres/CMakeLists.txt
@@ -225,6 +225,7 @@
CERES_TEST(corrector)
CERES_TEST(cost_function_to_functor)
CERES_TEST(dense_sparse_matrix)
+ CERES_TEST(dynamic_autodiff_cost_function)
CERES_TEST(evaluator)
CERES_TEST(gradient_checker)
CERES_TEST(gradient_checking_cost_function)
diff --git a/internal/ceres/dynamic_autodiff_cost_function_test.cc b/internal/ceres/dynamic_autodiff_cost_function_test.cc
new file mode 100644
index 0000000..5d743a7
--- /dev/null
+++ b/internal/ceres/dynamic_autodiff_cost_function_test.cc
@@ -0,0 +1,166 @@
+// Ceres Solver - A fast non-linear least squares minimizer
+// Copyright 2012 Google Inc. All rights reserved.
+// http://code.google.com/p/ceres-solver/
+//
+// Redistribution and use in source and binary forms, with or without
+// modification, are permitted provided that the following conditions are met:
+//
+// * Redistributions of source code must retain the above copyright notice,
+// this list of conditions and the following disclaimer.
+// * Redistributions in binary form must reproduce the above copyright notice,
+// this list of conditions and the following disclaimer in the documentation
+// and/or other materials provided with the distribution.
+// * Neither the name of Google Inc. nor the names of its contributors may be
+// used to endorse or promote products derived from this software without
+// specific prior written permission.
+//
+// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+// AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+// IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+// ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
+// LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+// CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+// SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+// INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+// CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+// ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+// POSSIBILITY OF SUCH DAMAGE.
+//
+// Author: thadh@gmail.com (Thad Hughes)
+// mierle@gmail.com (Keir Mierle)
+// sameeragarwal@google.com (Sameer Agarwal)
+
+#include "ceres/dynamic_autodiff_cost_function.h"
+
+#include <cstddef>
+
+#include "gtest/gtest.h"
+
+namespace ceres {
+namespace internal {
+
+// Takes 2 parameter blocks:
+// parameters[0] is size 10.
+// parameters[1] is size 5.
+// Emits 21 residuals:
+// A: i - parameters[0][i], for i in [0,10) -- this is 10 residuals
+// B: parameters[0][i] - i, for i in [0,10) -- this is another 10.
+// C: sum(parameters[0][i]^2 - 8*parameters[0][i]) + sum(parameters[1][i])
+class MyCostFunctor {
+ public:
+ template <typename T>
+ bool operator()(T const* const* parameters, T* residuals) const {
+ const T* params0 = parameters[0];
+ int r = 0;
+ for (int i = 0; i < 10; ++i) {
+ residuals[r++] = T(i) - params0[i];
+ residuals[r++] = params0[i] - T(i);
+ }
+
+ T c_residual(0.0);
+ for (int i = 0; i < 10; ++i) {
+ c_residual += pow(params0[i], 2) - T(8) * params0[i];
+ }
+
+ const T* params1 = parameters[1];
+ for (int i = 0; i < 5; ++i) {
+ c_residual += params1[i];
+ }
+ residuals[r++] = c_residual;
+ return true;
+ }
+};
+
+TEST(DynamicAutodiffCostFunctionTest, TestResiduals) {
+ vector<double> param_block_0(10, 0.0);
+ vector<double> param_block_1(5, 0.0);
+ DynamicAutoDiffCostFunction<MyCostFunctor, 3> cost_function(
+ new MyCostFunctor());
+ cost_function.AddParameterBlock(param_block_0.size());
+ cost_function.AddParameterBlock(param_block_1.size());
+ cost_function.SetNumResiduals(21);
+
+ // Test residual computation.
+ vector<double> residuals(21, -100000);
+ vector<double*> parameter_blocks(2);
+ parameter_blocks[0] = ¶m_block_0[0];
+ parameter_blocks[1] = ¶m_block_1[0];
+ EXPECT_TRUE(cost_function.Evaluate(¶meter_blocks[0],
+ residuals.data(),
+ NULL));
+ for (int r = 0; r < 10; ++r) {
+ EXPECT_EQ(1.0 * r, residuals.at(r * 2));
+ EXPECT_EQ(-1.0 * r, residuals.at(r * 2 + 1));
+ }
+ EXPECT_EQ(0, residuals.at(20));
+}
+
+TEST(DynamicAutodiffCostFunctionTest, TestJacobian) {
+ // Test the residual counting.
+ vector<double> param_block_0(10, 0.0);
+ for (int i = 0; i < 10; ++i) {
+ param_block_0[i] = 2 * i;
+ }
+ vector<double> param_block_1(5, 0.0);
+ DynamicAutoDiffCostFunction<MyCostFunctor, 3> cost_function(
+ new MyCostFunctor());
+ cost_function.AddParameterBlock(param_block_0.size());
+ cost_function.AddParameterBlock(param_block_1.size());
+ cost_function.SetNumResiduals(21);
+
+ // Prepare the residuals.
+ vector<double> residuals(21, -100000);
+
+ // Prepare the parameters.
+ vector<double*> parameter_blocks(2);
+ parameter_blocks[0] = ¶m_block_0[0];
+ parameter_blocks[1] = ¶m_block_1[0];
+
+ // Prepare the jacobian.
+ vector<vector<double> > jacobian_vect(2);
+ jacobian_vect[0].resize(21 * 10, -100000);
+ jacobian_vect[1].resize(21 * 5, -100000);
+ vector<double*> jacobian;
+ jacobian.push_back(jacobian_vect[0].data());
+ jacobian.push_back(jacobian_vect[1].data());
+
+ // Test jacobian computation.
+ EXPECT_TRUE(cost_function.Evaluate(parameter_blocks.data(),
+ residuals.data(),
+ jacobian.data()));
+
+ for (int r = 0; r < 10; ++r) {
+ EXPECT_EQ(-1.0 * r, residuals.at(r * 2));
+ EXPECT_EQ(+1.0 * r, residuals.at(r * 2 + 1));
+ }
+ EXPECT_EQ(420, residuals.at(20));
+ for (int p = 0; p < 10; ++p) {
+ // Check "A" Jacobian.
+ EXPECT_EQ(-1.0, jacobian_vect[0][2*p * 10 + p]);
+ // Check "B" Jacobian.
+ EXPECT_EQ(+1.0, jacobian_vect[0][(2*p+1) * 10 + p]);
+ jacobian_vect[0][2*p * 10 + p] = 0.0;
+ jacobian_vect[0][(2*p+1) * 10 + p] = 0.0;
+ }
+
+ // Check "C" Jacobian for first parameter block.
+ for (int p = 0; p < 10; ++p) {
+ EXPECT_EQ(4 * p - 8, jacobian_vect[0][20 * 10 + p]);
+ jacobian_vect[0][20 * 10 + p] = 0.0;
+ }
+ for (int i = 0; i < jacobian_vect[0].size(); ++i) {
+ EXPECT_EQ(0.0, jacobian_vect[0][i]);
+ }
+
+ // Check "C" Jacobian for second parameter block.
+ for (int p = 0; p < 5; ++p) {
+ EXPECT_EQ(1.0, jacobian_vect[1][20 * 5 + p]);
+ jacobian_vect[1][20 * 5 + p] = 0.0;
+ }
+ for (int i = 0; i < jacobian_vect[1].size(); ++i) {
+ EXPECT_EQ(0.0, jacobian_vect[1][i]);
+ }
+}
+
+} // namespace internal
+} // namespace ceres
