修改Centos默认ssh端口22

123

git clone https://github.com/openai/gymcd gympip install -e .

1

pip install -e .[all]

1234567

import gymenv = gym.make('CartPole-v0')env.reset()for _ in range(1000): env.render() env.step(env.action_space.sample()) # take a random actionenv.close()

1

gym\envs\classic_control\cartpole.py

123456

`register(` `id='CartPole-v1',` `entry_point='gym.envs.classic_control:CartPoleEnv',` `max_episode_steps=500,` `reward_threshold=475.0,` `)`

1

from gym.envs.classic_control.cartpole import CartPoleEnv

1

self.action_space = spaces.Box(low=-10, high=10, shape=(1,2))

1

self.observation_space = spaces.Discrete(2)

1

self.observation_space = spaces.MultiBinary(2)

1

self.observation_space = MultiDiscrete（）

12345

├── gym_test│   ├── env│   │   ├── env_guess_number.py│   │   ├── __init__.py│   └── __init__.py

12345678

from gym.envs.registration import registerregister( id='MYGUESSNUMBER-v0', entry_point='gym_test.env.env_guess_number:guess_number',)# gym_test.env是相对于项目名字的gym的路径 # env_guess_number是env_guess_number.py# guess_number 是类名

123

import gymimport gym_test.envenv = gym.make('MYGUESSNUMBER-v0')

12345678910

整个项目结构：├── gym_test│   ├── env│   │   ├── env.guess_number.py│   │   ├── __init__.py│   │   └── __pycache__│   │   └── __init__.cpython-37.pyc│   └── __init__.py├── README.md└── test.py

12345

from gym.envs.registration import registerregister( id='MYGUESSNUMBER-v0', entry_point='gym_test.env.env_guess_number:guess_number',)

1234567891011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465666768

import gymfrom gym import spacesfrom gym.utils import seedingimport numpy as npclass guess_number(gym.Env): """Hotter Colder The goal of hotter colder is to guess closer to a randomly selected guess_number After each step the agent receives an observation of: 0 - No guess yet submitted (only after reset) 1 - Guess is lower than the target 2 - Guess is equal to the target 3 - Guess is higher than the target The rewards is calculated as: (min(action, self.guess_number) + self.range) / (max(action, self.guess_number) + self.range) Ideally an agent will be able to recognise the 'scent' of a higher reward and increase the rate in which is guesses in that direction until the reward reaches its maximum """ def __init__(self): self.range = 1000 # +/- value the randomly select guess_number can be between self.bounds = 2000 # Action space bounds self.action_space = spaces.Box(low=np.array([-self.bounds]), high=np.array([self.bounds])) self.observation_space = spaces.Discrete(4) self.guess_number = 0 self.guess_count = 0 self.guess_max = 200 self.observation = 0 self.seed() self.reset() def seed(self, seed=None): self.np_random, seed = seeding.np_random(seed) return [seed] def step(self, action): # assert self.action_space.contains(action) if action < self.guess_number: self.observation = 1 elif action == self.guess_number: self.observation = 2 elif action > self.guess_number: self.observation = 3 reward = ((min(action, self.guess_number) + self.bounds) / (max(action, self.guess_number) + self.bounds)) ** 2 self.guess_count += 1 done = self.guess_count >= self.guess_max return self.observation, reward, done, {"guess_number": self.guess_number, "guesses": self.guess_count} def reset(self): self.guess_number = self.np_random.uniform(-self.range, self.range) print('guess number = ', self.guess_number) self.guess_count = 0 self.observation = 0 return self.observation

123456789101112

import gymimport gym_test.envenv = gym.make('MYGUESSNUMBER-v0')obs = env.reset()for step in range(10000): action = env.action_space.sample() print(action) obs, reward, done, info = env.step(action)