krshrimali

#include <iostream>
#include <cmath>
#include <vector>

int main() {
    std::vector<std::vector<double>> samples { {-7.5, -3.}, {-7.7, 3.1}, {-5.5, 3.}, {3., -2.5}, {3., 2.}, {7.3, 2.2}, {7.5, -2.} };
    std::cout << "Output format: a rem b is: std::remainder, output before PR, output after PR\n";
    for(auto vec: samples) {
        double a = vec[0], b = vec[1];

krshrimali

#include <iostream>
#include <cmath>
#include <vector>

int main() {
    std::vector<std::vector<int>> samples = { {-7, -3}, {-7, 3}, {-5, 3}, {3, -2}, {3, 2}, {7, 2}, {7, -2} };
    std::cout << "Output format: a rem b is: std::remainder, a % b in C++, output of torch impl, output modified impl\n";
    for(auto vec: samples) {
        int a = vec[0], b = vec[1];

krshrimali

1. Be comfortable with the library, learn how to use it and make some projects on it. For example, Django: make a website out of it, know what it does.
2. Be comfortable with "using" the library in C++ if it's written in that language. Python - C++ - try to use it in C++, OpenCV - C++.
3. Build the library from source. `pip3 install opencv-python, flask, django` - build everything from source.
  ```
  every library has 1000s of source codes! header files, (source) c++ files, test codes..
  ```
  while building - you'll have a lot of errors, try to solve it.
 4. Look at the source code "the functions you know something about".
 5. If you have any questions, reach out!
 6. Be more comfortable with the conventions followed in that library.

krshrimali

Visualizing rectangular 2D arrays in Python and Matplotlib the way you do with Matlab’s imagesc

Say you have a very rectangular 2D array arr, whose columns and rows correspond to very specific sampling locations x and y. That is, the arr[i, j] entry corresponds to some measurement taken at x[j] and y[i].

Matlab’s imagesc shows you this quite meaningfully:

x = linspace(-100, -10, 10);
y = [-8 -3];
data = randn(numel(y), numel(x));