insaneyilin

ffmpeg -ss 0 -t 3 -i input.mp4 \
    -vf "fps=10,scale=320:-1:flags=lanczos,split[s0][s1];[s0]palettegen[p];[s1][p]paletteuse" \
    -loop 0 output.gif

insaneyilin

# 保存到 ~/.gdbinit
python
import sys 
sys.path.insert(0, '/usr/share/gcc-4.8.2/python') # 这个路径以实际情况为准
from libstdcxx.v6.printers import register_libstdcxx_printers
register_libstdcxx_printers (None)
end

#                                                                                                        
#   STL GDB evaluators/views/utilities - 1.03

insaneyilin

def tile_along_axis(x, dim, n_tile):
    init_dim = x.size(dim)
    repeat_idx = [1] * x.dim()
    repeat_idx[dim] = n_tile
    x = x.repeat(*(repeat_idx))
    order_index = torch.tensor(
        torch.cat([init_dim * torch.arange(n_tile, device=x.device) + i for i in range(init_dim)]),
        dtype=torch.long, device=x.device)
    return torch.index_select(x, dim, order_index)

insaneyilin

#!/bin/bash
while true
do
    x=`xdotool getmouselocation | grep -oP '(?<=x:)\d+'`
    y=`xdotool getmouselocation | grep -oP '(?<=y:)\d+'`
    xdotool mousemove $x $y
    xdotool click 1
    sleep 1
done

insaneyilin

vector <string> Names {"Karl", "Martin", "Paul", "Jennie"};
vector <int> Score{45, 5, 14, 24};

std::vector<int> indices(Names.size());
 std::iota(indices.begin(), indices.end(), 0);
 std::sort(indices.begin(), indices.end(),
           [&](int A, int B) -> bool {
                return Score[A] < Score[B];
            });

insaneyilin

def masked_log_softmax(input, mask, dim=1):
    masked_input = input * mask.float()
    max_input = torch.max(masked_input, dim=dim, keepdim=True)[0]
    exps = torch.exp(masked_input - max_input)
    masked_exps = exps * mask.float()
    masked_sums = masked_exps.sum(dim, keepdim=True)
    zeros = (masked_sums == 0)
    masked_sums += zeros.float()
    masked_exps += 1e-6  # avoid zero input of log.
    return torch.log(masked_exps / masked_sums)

insaneyilin

# reference: https://github.com/rlabbe/Kalman-and-Bayesian-Filters-in-Python

import copy
import numpy as np
from scipy.linalg import block_diag
from filterpy.kalman import IMMEstimator
from filterpy.kalman import KalmanFilter
from filterpy.common import Q_discrete_white_noise

def make_cv_filter(dt):

insaneyilin

#pragma once

#include <cmath>

// https://en.wikipedia.org/wiki/Algorithms_for_calculating_variance#Welford's_online_algorithm
template <typename T>
class WelfordMeanStddevCalculator {
 public:
  WelfordMeanStddevCalculator() = default;
  virtual ~WelfordMeanStddevCalculator() = default;

insaneyilin

#define MODULE_NAME   "MY_LIB"
// __VA_ARGS__ 对应宏定义中参数列表的最后一个参数为省略号（也就是三个点）
// 加上##的作用在于，当可变参数的个数为0时，这里的##起到把前面多余的","去掉的作用,否则会编译出错
#define error_print(fmt, ...)  printf("[ERROR]["MODULE_NAME"](%s|%d)" fmt, __func__, __LINE__, ##__VA_ARGS__)

// 单个 # 用来把参数转换成字符串

// ## 是个粘合剂，将前后两部分粘合起来，也就是有“字符化”的意思。但是“##”不能随意粘合任意字符，必须是合法的C语言标示符。在单一的宏定义中，最多可以出现一次“#”或“##”预处理操作符。

insaneyilin

// The Hellinger distance measures the similarity of two probability
// distributions Reference: https://en.wikipedia.org/wiki/Hellinger_distance
double HellingerDistance(
    const std::vector<double>& histogram1,
    const std::vector<double>& histogram2) {
  const int histogram_size = histogram1.size();
  CHECK_GT(histogram_size, 0);
  CHECK_EQ(histogram_size, histogram2.size());
  const double mean1 =
      std::accumulate(histogram1.begin(), histogram1.end(), 0.0) /