Hitman-Go environments based on OpenAI gym.
- gym 0.14.0
- numpy
cd HitmanEnv
pip install -e .hitman-v0 environments is collective version of Hitman-Go environments. You can select maps by parameters (map_id)
import gym
import hitman_gym
from gym.envs.registration import register
map_id = 'simple' # 'simple', 'blue', 'yellow'
register(
id='hitman-v0',
entry_point='hitman_gym.envs:HitmanGO',
kwargs={'map_id': map_id}
)
env = gym.make('hitman-v0')
done = False
ep_reward = 0
while not done:
obs, reward, done, info = env.step(env.action_space.sample())
ep_reward += reward
env.close()Another versions of Hitman-Go do not need to register environments. By "gym.make" helps you use the hitman-go environments.
import gym
import hitman_gym
env = gym.make('hitman-v1') # hitman-v1: 'simple' , hitman-v2: 'blue', hitman-v3: 'yellow', hitman-v4: 'yellow_yr'
done = False
ep_reward = 0
while not done:
obs, reward, done, info = env.step(env.action_space.sample())
ep_reward += reward
env.close()python train.py --num_episodes 10000 --min_eps 0.01 --max_mem 1000 --map simplepython train2.py --num_episodes 10000 --min_eps 0.01 --max_mem 1000 --ver 1Sungkyunkwan University Reinforcement Learning Spring 2020 Youngrok Song Jiwung Hyun
RL OpenAI Gym Environment for Hitman Go
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent