karpathy

""" Trains an agent with (stochastic) Policy Gradients on Pong. Uses OpenAI Gym. """
import numpy as np
import cPickle as pickle
import gym

# hyperparameters
H = 200 # number of hidden layer neurons
batch_size = 10 # every how many episodes to do a param update?
learning_rate = 1e-4
gamma = 0.99 # discount factor for reward

JungeAlexander

# From http://stackoverflow.com/a/11158224

# Solution A - If the script importing the module is in a package
from .. import mymodule

# Solution B - If the script importing the module is not in a package
import os,sys,inspect
current_dir = os.path.dirname(os.path.abspath(inspect.getfile(inspect.currentframe())))
parent_dir = os.path.dirname(current_dir)
sys.path.insert(0, parent_dir) 

jrochkind

require 'date'

# Actually doesn't matter WHAT you choose as the epoch, it
# won't change the algorithm. Just don't change it after you
# have cached computed scores. Choose something before your first
# post to avoid annoying negative numbers. Choose something close
# to your first post to keep the numbers smaller. This is, I think,
# reddit's own epoch. 
$our_epoch = Time.local(2005, 12, 8, 7, 46, 43).to_time