mirror of
https://github.com/saymrwulf/stable-baselines3.git
synced 2026-05-18 21:30:19 +00:00
44f8218df0
* Split torch module code into torch_layers file * Updated reference to CNN * Change 'CxWxH' to 'CxHxW', as per common notion * Fix missing import in policies.py * Move PPOPolicy to OnlineActorCriticPolicy * Create OnPolicyRLModel from PPO, and make A2C and PPO inherit * Update A2C optimizer comment * Clean weight init scales for clarity * Fix A2C log_interval default parameter * Rename 'progress' to 'progress_remaining * Rename 'Models' to 'Algorithms' * Rename 'OnlineActorCriticPolicy' to 'ActorCriticPolicy' * Move static functions out from BaseAlgorithm * Move on/off_policy base algorithms to their own files * Add files for A2C/PPO * Fix docs * Fix pytype * Update documentation on OnPolicyAlgorithm * Add proper doctstring for on_policy rollout gathering * Add bit clarification on the mlppolicy/cnnpolicy naming * Move static function is_vectorized_policies to utils.py * Checking docstrings, pep8 fixes * Update changelog * Clean changelog * Remove policy warnings for sac/td3 * Add monitor_wrapper for OnPolicyAlgorithm. Clean tb logging variables. Add parameter keywords to OffPolicyAlgorithm super init Co-authored-by: Antonin RAFFIN <antonin.raffin@ensta.org>
62 lines
2 KiB
Python
62 lines
2 KiB
Python
import typing
|
|
import numpy as np
|
|
from gym import spaces
|
|
|
|
from stable_baselines3.common.vec_env.base_vec_env import VecEnv, VecEnvWrapper
|
|
from stable_baselines3.common.preprocessing import is_image_space
|
|
|
|
if typing.TYPE_CHECKING:
|
|
from stable_baselines3.common.type_aliases import GymStepReturn # noqa: F401
|
|
|
|
|
|
class VecTransposeImage(VecEnvWrapper):
|
|
"""
|
|
Re-order channels, from HxWxC to CxHxW.
|
|
It is required for PyTorch convolution layers.
|
|
|
|
:param venv: (VecEnv)
|
|
"""
|
|
|
|
def __init__(self, venv: VecEnv):
|
|
assert is_image_space(venv.observation_space), 'The observation space must be an image'
|
|
|
|
observation_space = self.transpose_space(venv.observation_space)
|
|
super(VecTransposeImage, self).__init__(venv, observation_space=observation_space)
|
|
|
|
@staticmethod
|
|
def transpose_space(observation_space: spaces.Box) -> spaces.Box:
|
|
"""
|
|
Transpose an observation space (re-order channels).
|
|
|
|
:param observation_space: (spaces.Box)
|
|
:return: (spaces.Box)
|
|
"""
|
|
assert is_image_space(observation_space), 'The observation space must be an image'
|
|
width, height, channels = observation_space.shape
|
|
new_shape = (channels, width, height)
|
|
return spaces.Box(low=0, high=255, shape=new_shape, dtype=observation_space.dtype)
|
|
|
|
@staticmethod
|
|
def transpose_image(image: np.ndarray) -> np.ndarray:
|
|
"""
|
|
Transpose an image or batch of images (re-order channels).
|
|
|
|
:param image: (np.ndarray)
|
|
:return: (np.ndarray)
|
|
"""
|
|
if len(image.shape) == 3:
|
|
return np.transpose(image, (2, 0, 1))
|
|
return np.transpose(image, (0, 3, 1, 2))
|
|
|
|
def step_wait(self) -> 'GymStepReturn':
|
|
observations, rewards, dones, infos = self.venv.step_wait()
|
|
return self.transpose_image(observations), rewards, dones, infos
|
|
|
|
def reset(self) -> np.ndarray:
|
|
"""
|
|
Reset all environments
|
|
"""
|
|
return self.transpose_image(self.venv.reset())
|
|
|
|
def close(self) -> None:
|
|
self.venv.close()
|