Implementation:Farama Foundation Gymnasium Wrappers Init

Knowledge Sources	Farama_Foundation_Gymnasium Gymnasium Docs
Domains	Reinforcement_Learning, Wrappers
Last Updated	2026-02-15 03:00 GMT

Overview

The public API index for the Gymnasium wrappers package, providing a centralized import point for all single-environment wrapper classes with lazy loading for framework-specific wrappers.

Description

This __init__.py module serves as the entry point for the gymnasium.wrappers package. It imports and re-exports all available wrapper classes organized by category:

Observation Wrappers: AtariPreprocessing, DelayObservation, DtypeObservation, DiscretizeObservation, FilterObservation, FlattenObservation, FrameStackObservation, GrayscaleObservation, TransformObservation, MaxAndSkipObservation, NormalizeObservation, AddRenderObservation, ResizeObservation, ReshapeObservation, RescaleObservation, TimeAwareObservation

Action Wrappers: ClipAction, DiscretizeAction, TransformAction, RescaleAction, StickyAction

Reward Wrappers: ClipReward, TransformReward, NormalizeReward

Common Wrappers: TimeLimit, Autoreset, PassiveEnvChecker, OrderEnforcing, RecordEpisodeStatistics

Rendering Wrappers: AddWhiteNoise, ObstructView, RenderCollection, RecordVideo, HumanRendering

Conversion Wrappers (lazy-loaded): ArrayConversion, JaxToNumpy, JaxToTorch, NumpyToTorch

The conversion wrappers (ArrayConversion, JaxToNumpy, JaxToTorch, NumpyToTorch) are loaded lazily via __getattr__ to avoid importing jax or torch at module load time. A _renamed_wrapper dictionary provides helpful error messages for deprecated wrapper names.

Usage

Import any wrapper directly from gymnasium.wrappers. This is the primary way users access wrappers in the Gymnasium API. The vector wrappers subpackage is also accessible as gymnasium.wrappers.vector.

Code Reference

Source Location

• Repository: Farama_Foundation_Gymnasium
• File: gymnasium/wrappers/__init__.py

Signature

# Module-level __getattr__ for lazy loading
def __getattr__(wrapper_name: str):
    """Load a wrapper by name, with lazy loading for framework-specific wrappers."""
    ...

Import

from gymnasium.wrappers import (
    ClipAction, RescaleAction, TransformAction, DiscretizeAction, StickyAction,
    NormalizeObservation, FrameStackObservation, TimeAwareObservation,
    NormalizeReward, TransformReward, ClipReward,
    TimeLimit, Autoreset, RecordEpisodeStatistics,
    RecordVideo, HumanRendering,
)
# Lazy-loaded:
from gymnasium.wrappers import ArrayConversion, JaxToNumpy, JaxToTorch, NumpyToTorch

I/O Contract

Exported Symbols

Category	Wrappers
Observation	AtariPreprocessing, DelayObservation, DtypeObservation, DiscretizeObservation, FilterObservation, FlattenObservation, FrameStackObservation, GrayscaleObservation, TransformObservation, MaxAndSkipObservation, NormalizeObservation, AddRenderObservation, ResizeObservation, ReshapeObservation, RescaleObservation, TimeAwareObservation
Action	ClipAction, DiscretizeAction, TransformAction, RescaleAction, StickyAction
Reward	ClipReward, TransformReward, NormalizeReward
Common	TimeLimit, Autoreset, PassiveEnvChecker, OrderEnforcing, RecordEpisodeStatistics
Rendering	AddWhiteNoise, ObstructView, RenderCollection, RecordVideo, HumanRendering
Conversion	ArrayConversion, JaxToNumpy, JaxToTorch, NumpyToTorch

Renamed Wrappers

Old Name	New Name
AutoResetWrapper	Autoreset
FrameStack	FrameStackObservation
PixelObservationWrapper	AddRenderObservation
VectorListInfo	vector.DictInfoToList

Usage Examples

import gymnasium as gym
from gymnasium.wrappers import RescaleAction

base_env = gym.make("Hopper-v4")
base_env.action_space
# Box(-1.0, 1.0, (3,), float32)

wrapped_env = RescaleAction(base_env, min_action=0, max_action=1)
wrapped_env.action_space
# Box(0.0, 1.0, (3,), float32)

# Wrappers can be chained
from gymnasium.wrappers import NormalizeObservation, NormalizeReward
env = gym.make("CartPole-v1")
env = NormalizeObservation(env)
env = NormalizeReward(env)

Related Pages

• Environment:Farama_Foundation_Gymnasium_Python_3_10_Runtime