matthiasplappert

import numpy as np
import gym


env = gym.make('FetchReach-v0')
obs = env.reset()
done = False

def policy(observation, desired_goal):
    # Here you would implement your smarter policy. In this case,