Kobzol

import collections
import dataclasses
import random
from typing import Optional, Tuple

import cv2
import numpy as np
import numpy.random
import tensorflow as tf
from tqdm import tqdm

Kobzol

Bulding TF 2.2 with TensorRT 7.1 from source

This is a short tutorial for installing Tensorflow 2.2 with Cuda 10.2, CuDNN 8 and TensorRT 7.1 on Ubuntu 18.04 with Python 3.6 and GCC 7.5. It works on my machine as of 23. 07. 2020 :) However, if you're reading this, the guide is probably already outdated.

If you have the exact same specification as I do, you can try your luck with my prebuilt pip wheel: https://mega.nz/file/MAozBYhI#8Tp7W8t-X5sBuWCVEviTN5quQnNNwaVMv90OAnEvGD0

Recently I got hold of Nvidia Xavier NX, which has JetPack 4.4 with Tensorflow 2.2 and TensortRT 7.1. I wanted to have the same environment on my desktop/server, but sadly, Tensorflow does not support CUDA 10.2, TensorRT 7.1 and CuDNN 8 by default

Kobzol

#include <zmq.hpp>
#include <sys/epoll.h>
#include <iostream>

zmq_msg_t msg()
{
    zmq_msg_t msg;
    zmq_msg_init_size(&msg, 5);
    std::memcpy(zmq_msg_data(&msg), "Hello", 5);
    return msg;

Kobzol

import msgpack

BYTES = []


def parse_messages(stream):
    data = stream.read(8)
    if len(data) < 8:
        return []
    chunk_count = int.from_bytes(data, "little", signed=False)

Kobzol

#include <zmq.hpp>
#include <sys/epoll.h>
#include <iostream>

int main()
{
    void* ctx = zmq_ctx_new();
    void* socket = zmq_socket(ctx, ZMQ_PUSH);
    std::cerr << zmq_connect(socket, "tcp://127.0.0.1:5555") << std::endl;

Kobzol

str_hodnota = function(v, p)
{
  sum(v * p);
}
rozptyl = function(v, p)
{
  str_hodnota(v ** 2, p) - str_hodnota(v, p) ** 2
}
var_koeficient = function(v, p)
{

Kobzol

str_hodnota = function(v, p)
{
  sum(v * p);
}
rozptyl = function(v, p)
{
  str_hodnota(v ** 2, p) - str_hodnota(v, p) ** 2
}
var_koeficient = function(v, p)
{

Kobzol

void generate(int index, int wordSize, int alphabetSize, char* result, char base = 0)
{
	for (int i = wordSize - 1; i >= 0; i--)
	{
		int toSkip = std::pow(alphabetSize, i);
		char c = index / toSkip;
		result[wordSize - i - 1] = base + c;
		index = index % toSkip;
	}
}

Kobzol

#include <iostream>

class A
{
public:
    void fn() & {
        std::cout << "A" << std::endl;
    }
    void fn() && {
        std::cout << "B" << std::endl;

Kobzol

template <typename T, unsigned int S>
void convolution(cv::Mat& original, cv::Mat& resultImg, T kernel[S][S])
{
	original.copyTo(resultImg);

	int offset = S / 2;
	T scale = 0;
	for (int i = 0; i < S; i++)
	{
		for (int j = 0; j < S; j++)