| #!/usr/bin/env python3 | |
| """ | |
| ll-cma.py: Solve LunarLander-v2 using CMA-ES. | |
| This code solves the LunarLander-v2 environment from the OpenAI gym using CMA-ES. | |
| You need to pip-install the packages numpy, gym[box2d] and cma in order for it to work. | |
| It doesn't seem to solve the environment in every run, but if it does, it |
This is a high-level description of my Halite III bot. At the time of writing, a few days before the end of the competition, it has a rating of slightly over 74, which puts it at rank ~115.
The bot is based on Value Iteration and Uniform Cost Search for mining halite and returning it to the nearest dropoff. Dropoffs are not pre-planned -- a ship
| defmodule FizzBuzz do | |
| @moduledoc """ | |
| FizzBuzz outputs integers interspersed with configurable words | |
| at configurable intervals. | |
| """ | |
| def fizzbuzz(count, words \\ [{3, "Fizz"}, {5, "Buzz"}]) do | |
| words | |
| |> Enum.map(&make_word_stream/1) | |
| |> Stream.zip |
| (ns clj-fizzbuzz.core | |
| (:gen-class)) | |
| (defn hitword | |
| [word n] | |
| (cycle (cons word (repeat (dec n) nil)))) | |
| (defn fizzbuzz | |
| [x] | |
| (let [fizzes (hitword "Fizz" 3) |
| {- | |
| "dependencies": { | |
| "eeue56/elm-flat-matrix": "3.0.2 <= v < 4.0.0", | |
| "elm-lang/core": "5.1.1 <= v < 6.0.0", | |
| "elm-lang/html": "2.0.0 <= v < 3.0.0", | |
| "elm-lang/svg": "2.0.0 <= v < 3.0.0" | |
| }, | |
| "elm-version": "0.18.0 <= v < 0.19.0" | |
| -} |
| import math | |
| import tensorflow as tf | |
| import numpy as np | |
| import tensorflow.contrib.slim as slim | |
| from scipy import misc | |
| class IMR(object): | |
| def __init__(self, sess, channels = 3): | |
| self.sess = sess | |
| self.num_channels= channels |
| import gym | |
| import numpy as np | |
| # This is a straightforwad implementation of SARSA for the FrozenLake OpenAI | |
| # Gym testbed. I wrote it mostly to make myself familiar with the OpenAI gym; | |
| # the SARSA algorithm was implemented pretty much from the Wikipedia page alone. | |
| env = gym.make("FrozenLake-v0") | |
| def choose_action(observation): |
| package main | |
| // This program takes a .zim-file and dumps all contained articles below the | |
| // current directory. I tested it with a Wikipedia snapshot; that did not | |
| // contain deleted articles or LinkTargetEntrys, so I'm unsure how to handle | |
| // those, for now I'm ignoring them. | |
| // | |
| // Redirects are handled by simply writing out the page pointed to by the | |
| // redirect. IPFS deduplication should take care of it, so I think this is the | |
| // most economical solution, even better than writing out small HTML files with |
| #!/usr/bin/env ruby | |
| require 'securerandom' | |
| # Generates a random password made up of NUM_SYLLABLES koremutake syllables. | |
| # See http://shorl.com/koremutake for more on koremutake. | |
| NUM_SYLLABLES=14 | |
| SYLLABLES = %w{ |
| # This entry is based on Prediction by Partial Matching, a method that is used | |
| # in data compression. | |
| import random | |
| from collections import defaultdict | |
| CONTEXT_SIZE=32 | |
| if input == "": | |
| # Initialize data structure |