ahmadyan

x = tf.Variable([1., 3.], tf.float32)
# Rosenbrock function
y = tf.add(tf.pow(tf.subtract(1.0, x[0]), 2.0), 
           tf.multiply(100.0, tf.pow(tf.subtract(x[1],tf.pow(x[0], 2.0)), 2.0)), 'y')

dx = tf.gradients(y, x)[0]
val = np.array([1, 3], dtype=np.float32)

with tf.Session() as sess:
    sess.run(tf.global_variables_initializer())

ahmadyan

# Build a function that also returns gradients using autograd.
rosenbrock_with_grad = autograd.value_and_grad(rosenbrock)

# Optimize using conjugate gradients from scipy
result = scipy.optimize.minimize(rosenbrock_with_grad, x0=np.array([0.0, 0.0]), jac=True, method='CG')

ahmadyan

x = [Variable(FloatTensor([1]), requires_grad=True), 
     Variable(FloatTensor([3]), requires_grad=True)]
y = rosenbrock(x)
y.backward()
print(y, x[0].grad, x[1].grad)

#Expect to see 400, -800, 400 as a result

ahmadyan

auto rosenbrockCostFunction = new ceres::AutoDiffCostFunction<Rosenbrock, 1, 1, 1>(new Rosenbrock());
constexpr double x = 1, y = 3;
parameters[0][0] = x;
parameters[1][0] = y;

rosenbrockCostFunction->Evaluate(parameters, &residuals, jacobians);
cout << residuals << " df/dx=" << jacobians[0][0] << " df/dy" << jacobians[1][0] << "\n";

ahmadyan

def rosenbrock(x):
   return 100*(x[1] - x[0]**2)**2 + (1 - x[0])**2

ahmadyan

// Rosenbrock function, a = 1.0, b = 100
class Rosenbrock {
public:
  template <typename T>
  bool operator()(const T *const x, const T *const y, T *cost) const {
    cost[0] = (T(1.0) - x[0]) * (T(1.0) - x[0]) +
              T(100.0) * (y[0] - x[0] * x[0]) * (y[0] - x[0] * x[0]);
    return true;
  }
};

ahmadyan

Jet operator*(Jet& f, Jet& g) {
  return Jet(f.a * g.a,            // The scalar part
             f.a * g.v + f.v * g.a // The infinitesimal vector part
            );
}

ahmadyan

// Ceres Solver - A fast non-linear least squares minimizer
// Copyright 2015 Google Inc. All rights reserved.
// http://ceres-solver.org/
//
// 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,

ahmadyan

#include <benchmark/benchmark.h>
#include <algorithm>
 
#include <iostream>
#include <vector>
#include <random>
#include <list>
#include <array>
#include <queue>  
#include <deque>

ahmadyan

#include <iostream>
#include <vector>
#include <random>
#include <list>
#include <array>
#include <queue>  
#include <deque>
using namespace std;

class Graph{