saberbud/ceres-solver
1
0
Fork 0
mirror of https://github.com/ceres-solver/ceres-solver.git synced 2026-08-31 01:20:38 +08:00
Code Issues Packages Projects Releases Wiki Activity

Reduce the number of individual PRNG instances

Browse Source 
Use same instance of a PRNG throughout by passing it to methods and
functions as an argument to generate random numbers without breaking the
sequence.

Change-Id: Ib024bbc1ea2d14e4b9afb71857856a5fb77b1667
This commit is contained in:
Sergiu Deitsch
2022-08-08 21:06:22 +02:00
committed by Sameer Agarwal
parent 79e403b15c
commit f1dfac8cd6
23 changed files with 186 additions and 140 deletions
Download Patch File Download Diff File Expand all files Collapse all files
internal/ceres/gradient_checking_cost_function_test.cc
+14 -12
View File
@@ -1,5 +1,5 @@
// Ceres Solver - A fast non-linear least squares minimizer
// Copyright 2015 Google Inc. All rights reserved.
// Copyright 2022 Google Inc. All rights reserved.
// http://ceres-solver.org/
//
// Redistribution and use in source and binary forms, with or without
@@ -69,15 +69,15 @@ class TestTerm : public CostFunction {
public:
// The constructor of this function needs to know the number
// of blocks desired, and the size of each block.
TestTerm(int arity, int const* dim) : arity_(arity) {
std::mt19937 prng;
std::uniform_real_distribution<double> uniform01(-1.0, 1.0);
template <class UniformRandomFunctor>
TestTerm(int arity, int const* dim, UniformRandomFunctor&& randu)
: arity_(arity) {
// Make 'arity' random vectors.
a_.resize(arity_);
for (int j = 0; j < arity_; ++j) {
a_[j].resize(dim[j]);
for (int u = 0; u < dim[j]; ++u) {
a_[j][u] = uniform01(prng);
a_[j][u] = randu();
}
}
@@ -139,11 +139,12 @@ TEST(GradientCheckingCostFunction, ResidualsAndJacobiansArePreservedTest) {
// Make a random set of blocks.
vector<double*> parameters(arity);
std::mt19937 prng;
std::uniform_real_distribution<double> uniform01(-1.0, 1.0);
std::uniform_real_distribution<double> distribution(-1.0, 1.0);
auto randu = [&prng, &distribution] { return distribution(prng); };
for (int j = 0; j < arity; ++j) {
parameters[j] = new double[dim[j]];
for (int u = 0; u < dim[j]; ++u) {
parameters[j][u] = uniform01(prng);
parameters[j][u] = randu();
}
}
@@ -162,7 +163,7 @@ TEST(GradientCheckingCostFunction, ResidualsAndJacobiansArePreservedTest) {
const double kRelativeStepSize = 1e-6;
const double kRelativePrecision = 1e-4;
TestTerm<-1, -1> term(arity, dim);
TestTerm<-1, -1> term(arity, dim, randu);
GradientCheckingIterationCallback callback;
auto gradient_checking_cost_function =
CreateGradientCheckingCostFunction(&term,
@@ -196,11 +197,12 @@ TEST(GradientCheckingCostFunction, SmokeTest) {
// Make a random set of blocks.
vector<double*> parameters(arity);
std::mt19937 prng;
std::uniform_real_distribution<double> uniform01(-1.0, 1.0);
std::uniform_real_distribution<double> distribution(-1.0, 1.0);
auto randu = [&prng, &distribution] { return distribution(prng); };
for (int j = 0; j < arity; ++j) {
parameters[j] = new double[dim[j]];
for (int u = 0; u < dim[j]; ++u) {
parameters[j][u] = uniform01(prng);
parameters[j][u] = randu();
}
}
@@ -218,7 +220,7 @@ TEST(GradientCheckingCostFunction, SmokeTest) {
// Should have one term that's bad, causing everything to get dumped.
LOG(INFO) << "Bad gradient";
{
TestTerm<1, 2> term(arity, dim);
TestTerm<1, 2> term(arity, dim, randu);
GradientCheckingIterationCallback callback;
auto gradient_checking_cost_function =
CreateGradientCheckingCostFunction(&term,
@@ -238,7 +240,7 @@ TEST(GradientCheckingCostFunction, SmokeTest) {
// The gradient is correct, so no errors are reported.
LOG(INFO) << "Good gradient";
{
TestTerm<-1, -1> term(arity, dim);
TestTerm<-1, -1> term(arity, dim, randu);
GradientCheckingIterationCallback callback;
auto gradient_checking_cost_function =
CreateGradientCheckingCostFunction(&term,