mirror of
https://github.com/wassname/DeepRL.git
synced 2026-08-23 11:50:41 +08:00
57 lines
1.6 KiB
Python
57 lines
1.6 KiB
Python
#######################################################################
|
|
# Copyright (C) 2017 Shangtong Zhang(zhangshangtong.cpp@gmail.com) #
|
|
# Permission given to modify the code as long as you keep this #
|
|
# declaration at the top #
|
|
#######################################################################
|
|
import gym
|
|
import sys
|
|
import numpy as np
|
|
import cv2
|
|
|
|
class BasicTask:
|
|
def transfer_state(self, state):
|
|
return state
|
|
|
|
def reset(self):
|
|
return self.transfer_state(self.env.reset())
|
|
|
|
def step(self, action):
|
|
next_state, reward, done, info = self.env.step(action)
|
|
next_state = self.transfer_state(next_state)
|
|
return next_state, np.sign(reward), done, info
|
|
|
|
class MountainCar(BasicTask):
|
|
name = 'MountainCar-v0'
|
|
success_threshold = -110
|
|
|
|
def __init__(self):
|
|
self.env = gym.make(self.name)
|
|
self.env._max_episode_steps = sys.maxsize
|
|
|
|
class CartPole(BasicTask):
|
|
name = 'CartPole-v0'
|
|
success_threshold = 195
|
|
|
|
def __init__(self):
|
|
self.env = gym.make(self.name)
|
|
|
|
class LunarLander(BasicTask):
|
|
name = 'LunarLander-v2'
|
|
success_threshold = 200
|
|
|
|
def __init__(self):
|
|
self.env = gym.make(self.name)
|
|
|
|
class PixelAtari(BasicTask):
|
|
width = 84
|
|
height = 84
|
|
success_threshold = 1000
|
|
|
|
def __init__(self, name):
|
|
self.env = gym.make(name)
|
|
|
|
def transfer_state(self, state):
|
|
img = (state[:, :, 0] * 0.299 + state[:, :, 1] * 0.587 + state[:, :, 2] * 0.114) / 255.0
|
|
img = cv2.resize(img, (self.width, self.height))
|
|
return np.reshape(img, (1, self.width, self.height))
|