jdhao

import numpy as np
import numba

class MaxPooling(object):

    def __init__(self, X, kernel_size=(2,2), stride=(2,2)):

        if len(X.shape) != 4:
            raise ValueError("Input must have be a tensor of shape N*C*H*W!")

jdhao

import cv2
import numpy as np


def main():
    img = cv2.imread("test_image.jpg")
    
    # assume coord is a list with 8 float values, the points of the rectangle area should
    # have be clockwise
    x1, y1, x2, y2, x3, y3, x4, y4 = coord

jdhao

How do I grep unicode characters using its code point?

For example, if we want to grep a(unicode code point is u+0041), we can use the following trick:

grep "$(printf '\u0041')" my_file.txt

References

jdhao

{
    "shell_cmd": "g++ -std=c++11 -Wall \"${file}\" -o \"${file_path}/${file_base_name}\" && \"${file_path}/${file_base_name}\"",
    "file_regex": "^(..[^:]*):([0-9]+):?([0-9]+)?:? (.*)$",
    "working_dir": "${file_path}",
    "selector": "source.c++, source.cpp, source.cc, source.cxx",

    "variants":
    [
        {
            "name": "Run in Terminal",

jdhao

echo "Installing dependency packages..."
# install packages which can be found by yum
yum -y install libcurl-devel libX11-devel libXt-devel libXinerama-devel libpng-devel

# download and install packages which are not in yum repo
wget ftp://ftp.pbone.net/mirror/ftp5.gwdg.de/pub/opensuse/repositories/home:/Kenzy:/modified:/C7/CentOS_7/x86_64/imlib2-1.4.6-2.1.x86_64.rpm
wget ftp://ftp.pbone.net/mirror/ftp5.gwdg.de/pub/opensuse/repositories/home:/Kenzy:/modified:/C7/CentOS_7/x86_64/imlib2-devel-1.4.6-2.1.x86_64.rpm
wget https://jaist.dl.sourceforge.net/project/libjpeg-turbo/1.5.1/libjpeg-turbo-official-1.5.1.x86_64.rpm
yum --nogpgcheck localinstall imlib2-1.4.6-2.1.x86_64.rpm imlib2-devel-1.4.6-2.1.x86_64.rpm libjpeg-turbo-official-1.5.1.x86_64.rpm

jdhao

Special characters

expression	Description
.	Match any character
^	Match line begin
$	Match line end
*	Match previous RE 0 or more times greedily
*?	Match previous RE 0 or more times non-greedily
+	Match previous RE 1 or more times greedily

jdhao

" Remove trailing white space, see https://vi.stackexchange.com/a/456/15292
function utils#StripTrailingWhitespaces() abort
    let l:save = winsaveview()
    " vint: next-line -ProhibitCommandRelyOnUser -ProhibitCommandWithUnintendedSideEffect
    keeppatterns %s/\v\s+$//e
    call winrestview(l:save)
endfunction

" Create command alias safely, see https://bit.ly/2ImFOpL.
" The following two functions are taken from answer below on SO:

jdhao

echo "Downloading gcc source files..."
curl https://ftp.gnu.org/gnu/gcc/gcc-5.4.0/gcc-5.4.0.tar.bz2 -O

echo "extracting files..."
tar xvfj gcc-5.4.0.tar.bz2

echo "Installing dependencies..."
yum -y install gmp-devel mpfr-devel libmpc-devel

echo "Configure and install..."

jdhao

Conversion between bytes and string in Python 3

To convert Unicode string to bytes object, you can use two methods:

• 'hello'.encode('utf-8')
• bytes('hello', encoding='utf-8')

To convert bytes back to Unicode string, you can use two methods:

• b'\xe4\xbd\xa0\xe5\xa5\xbd'.decode('utf-8')

jdhao

"""
in this script, we calculate the image per channel mean and standard
deviation in the training set, do not calculate the statistics on the
whole dataset, as per here http://cs231n.github.io/neural-networks-2/#datapre
"""

import numpy as np
from os import listdir
from os.path import join, isdir
from glob import glob