unixpickle

Overview

In this document, I will describe how I managed to get unlimited stars and coins in the Android game "Lily's Garden". In short, here are the steps:

• Use the adb backup feature to extract all of the game's data
• Extract the Android backup into a tar file
• Modify the file which stores the number of coins and stars
• Re-sign the coins/stars field using a reverse-engineered HMAC key
• Convert the tar file back to an Android backup

unixpickle

import numpy as np
from tqdm.auto import tqdm

def is_bitonic(xs):
    results = []
    for seq in xs:
        prev = seq[0]
        direction = 0
        num_changes = 0
        for x1 in seq:

unixpickle

Benchmarking global reductions with varying numbers of SMs and warps per SM. We find interesting facts, like that using one warp across more SMs is more efficient than more warps across a single SM.

These were computed on an NVIDIA Titan X GPU.

Code for the benchmark is here and the kernel is here.

1 warps

unixpickle

import torch
import torch.nn.functional as F


def maml_grad(model, inputs, outputs, lr, batch=1):
    """
    Update a model's gradient using MAML.

    The gradient will point in the direction that
    improves the total loss across all inner-loop

unixpickle

unixpickle

// Solves this problem: http://1mage.us/1339.

package main

import (
	"fmt"
	"math"
	"math/rand"
	"sort"

unixpickle

"""
Compute the closed-form characteristic polynomial of a 4x4 matrix.
"""

import sympy
syms = sympy.symbols('a b c d e f g h i j k l m n o p')
mat = sympy.Matrix([[syms[i+j*4] for i in range(4)] for j in range(4)])
print(sympy.collect((mat - sympy.eye(4)*sympy.symbols('x')).det(), 'x'))

# a*f*k*p - a*f*l*o - a*g*j*p + a*g*l*n + a*h*j*o - a*h*k*n - b*e*k*p + b*e*l*o + b*g*i*p - b*g*l*m - b*h*i*o + b*h*k*m + c*e*j*p - c*e*l*n - c*f*i*p + c*f*l*m + c*h*i*n - c*h*j*m - d*e*j*o + d*e*k*n + d*f*i*o - d*f*k*m - d*g*i*n + d*g*j*m + x**4 + x**3*(-a - f - k - p) + x**2*(a*f + a*k + a*p - b*e - c*i - d*m + f*k + f*p - g*j - h*n + k*p - l*o) + x*(-a*f*k - a*f*p + a*g*j + a*h*n - a*k*p + a*l*o + b*e*k + b*e*p - b*g*i - b*h*m - c*e*j + c*f*i + c*i*p - c*l*m - d*e*n + d*f*m - d*i*o + d*k*m - f*k*p + f*l*o + g*j*p - g*l*n - h*j*o + h*k*n)

unixpickle

#!/bin/bash
#
# Simply pass in your plex library database file and have some fun!
# NOTE: your DB file is probably called "com.plexapp.plugins.library.db"
#

if [ $# != 1 ]; then
    echo "plexdblist - lists the media items in a plex DB file."
	echo "Usage: plexdblist <database>"
	exit 1

unixpickle

#!/usr/bin/perl

use strict;

sub arrcount {
	my ($arr) = @_;
	my $c = 0;
	foreach my $str (@{$arr}) {
		++$c;
	}

unixpickle

import numpy as np

rows = 3
cols = 6

# Create all 2^(rows*cols) combinations of egg matrices.
all_combos = np.zeros([1] * (rows * cols + 2), dtype=np.float32)
for i in range(rows * cols):
    m = np.zeros([2, rows * cols], dtype=np.float32)
    m[1, i] = 1