cherry.wrappers¶
cherry.wrappers.base_wrapper.Wrapper
¶
Description¶
This class allows to chain Environment Wrappers while still being able to access the properties of wrapped wrappers.
Example¶
env = gym.make('MyEnv-v0')
env = cherry.wrappers.Logger(env)
env = cherry.wrappers.Runner(env)
env.log('asdf', 23) # Uses log() method from cherry.wrappers.Logger.
cherry.wrappers.runner_wrapper.Runner
¶
Description¶
Helps collect transitions, given a get_action function.
Example¶
env = MyEnv()
env = Runner(env)
replay = env.run(lambda x: policy(x), steps=100)
# or
replay = env.run(lambda x: policy(x), episodes=5)
run(self, get_action, steps = None, episodes = None, render = False)
¶
Description¶
Runner wrapper's run method.
Info
Either use the steps OR the episodes argument.
Arguments¶
get_action(function) - Given a state, returns the action to be taken.steps(int, optional, default=None) - The number of steps to be collected.episodes(int, optional, default=None) - The number of episodes to be collected.
cherry.wrappers.torch_wrapper.Torch
¶
This wrapper converts * actions from Tensors to numpy, * states from lists/numpy to Tensors.
Examples:
action = Categorical(Tensor([1, 2, 3])).sample() env.step(action)
__init__(self, env, device = None, env_device = None)
special
¶
cherry.wrappers.timestep_wrapper.AddTimestep
¶
Adds a timestep information to the state input.
Modified from Ilya Kostrikov's implementation:
https://github.com/ikostrikov/pytorch-a2c-ppo-acktr/
__init__(self, env = None)
special
¶
cherry.wrappers.action_space_scaler_wrapper.ActionSpaceScaler
¶
Scales the action space to be in the range (-clip, clip).
Adapted from Vitchyr Pong's RLkit: https://github.com/vitchyr/rlkit/blob/master/rlkit/envs/wrappers.py#L41
__init__(self, env, clip = 1.0)
special
¶
Soon Deprecated¶
Info
The following wrappers will soon be deprecated because they are available in gym.