kvfrans

kvfrans

kvfrans

import numpy as np
import random
import json
import scipy.stats as ss
import multiprocessing as mp
import time
import sys

import torch
import torch.nn as nn

kvfrans

import numpy as np
import argparse
import tensorflow as tf
import time
import random

# This is an improved version of Q_Basic. It has experience replay, and remembers previous transitions to train on again.
# In order to fix non-convergence problems, I manually put a reward of -200 when failing to reach 200 timesteps, and I
# run 10 supervised training updates after each episode.

kvfrans

import tensorflow as tf
import numpy as np
import random
import gym
import math
import matplotlib.pyplot as plt


def softmax(x):
    e_x = np.exp(x - np.max(x))

kvfrans

import gym
import numpy as np

def run_episode(env, parameters):
    observation = env.reset()
    totalreward = 0
    counter = 0
    for _ in xrange(200):
        # env.render()
        action = 0 if np.matmul(parameters,observation) < 0 else 1

kvfrans

import gym
import numpy as np

def run_episode(env, parameters):
    observation = env.reset()
    totalreward = 0
    while True:
        env.render()
        action = 0 if np.matmul(parameters,observation) < 0 else 1
        observation, reward, done, info = env.step(action)