Class AntEnvV5
- Namespace
- Gym.Environments.Envs.MuJoCo
- Assembly
- Gym.Environments.dll
Ant-v5: a four-legged ant (a torso and four two-link legs on a free joint) driven by 8 hip and ankle torques in [-1, 1]. The reward pays the main body's forward velocity and staying healthy, and charges the squared torques and the clipped contact forces; a non-finite state or a torso height outside [0.2, 1.0] ends the episode. Byte for byte what Gymnasium's Ant-v5 computes on MuJoCo 4.0.0 (DR-61, DR-64).
[PyName("AntEnv")]
public class AntEnvV5 : MujocoEnv, IEnv, IDisposable, IDynamicAttributes
- Inheritance
-
AntEnvV5
- Implements
- Inherited Members
- Extension Methods
Remarks
Observation (float64): the positions without the root's x and y (unless asked for), the velocities, then the
contact forces of every body but the world clipped to contact_force_range (unless
include_cfrc_ext_in_observation=False), 105 values by default. Reward:
x_velocity * forward_reward_weight + healthy_reward * is_healthy - ctrl_cost_weight * sum(action**2) -
contact_cost_weight * sum(clip(cfrc_ext)**2), the velocity measured on main_body's frame. Info:
x_position, y_position, distance_from_origin (of the root joint), x_velocity,
y_velocity, reward_forward, reward_ctrl, reward_contact, reward_survive; the
reset's info holds the first three.
Not thread-safe. Reference: gymnasium/envs/mujoco/ant_v5.py.
Its info reads typed (DR-91): AntInfoV5 info = step.Info; (AntInfoV5), with the dict's own values still behind the indexer.
Constructors
AntEnvV5(AntEnvV5Options)
Builds the instance from its options record: the properties set, the constructor's defaults for the rest.
public AntEnvV5(AntEnvV5Options options)
Parameters
optionsAntEnvV5OptionsThe options.
Exceptions
- ArgumentNullException
Thrown when
optionsis null.
AntEnvV5(RenderMode, string, int, IReadOnlyDictionary<string, object?>?, double, double, double, double, object?, bool, (double Low, double High)?, (double Low, double High)?, double, bool, bool, int, int, int?, string?, int, IReadOnlyDictionary<int, bool>?, EnvViewerFactory?)
Builds the env with the render mode given as the RenderMode enum (DR-84, DR-93), as
AntEnvV5(string, int, IReadOnlyDictionary<string, object?>?, double, double, double, double, object?, bool, (double Low, double High)?, (double Low, double High)?, double, bool, bool, string?, int, int, int?, string?, int, IReadOnlyDictionary<int, bool>?, EnvViewerFactory?)
builds it from the mode's string: new AntEnvV5(RenderMode.RgbArray) builds what
new AntEnvV5(renderMode: "rgb_array") builds, the member reaching the env as Gymnasium's string
(Name(RenderMode)), which its RenderMode property then holds.
public AntEnvV5(RenderMode renderMode, string xmlFile = "ant.xml", int frameSkip = 5, IReadOnlyDictionary<string, object?>? defaultCameraConfig = null, double forwardRewardWeight = 1, double ctrlCostWeight = 0.5, double contactCostWeight = 0.0005, double healthyReward = 1, object? mainBody = null, bool terminateWhenUnhealthy = true, (double Low, double High)? healthyZRange = null, (double Low, double High)? contactForceRange = null, double resetNoiseScale = 0.1, bool excludeCurrentPositionsFromObservation = true, bool includeCfrcExtInObservation = true, int width = 480, int height = 480, int? cameraId = null, string? cameraName = null, int maxGeom = 1000, IReadOnlyDictionary<int, bool>? visualOptions = null, EnvViewerFactory? viewerFactory = null)
Parameters
renderModeRenderModeThe render mode, which the env must list in its
metadata["render_modes"]as for the string; it behaves exactly as its Gymnasium string does in the other constructor.xmlFilestringThe model: a bundled file name (
xml_file, DR-63).frameSkipintPhysics steps per env step (
frame_skip).defaultCameraConfigIReadOnlyDictionary<string, object>The free camera's settings (
default_camera_config); null is Gymnasium's default,{"distance": 4.0}, and an empty dictionary leaves MuJoCo's camera as it is.forwardRewardWeightdoubleThe weight of the forward velocity (
forward_reward_weight; Gymnasium's default is the int 1).ctrlCostWeightdoubleThe weight of the squared torques (
ctrl_cost_weight).contactCostWeightdoubleThe weight of the squared clipped contact forces (
contact_cost_weight).healthyRewarddoubleThe reward for each healthy step (
healthy_reward).mainBodyobjectThe body whose motion is the forward velocity (
main_body): its index (an integer, a bool or a 0-d integer array) or its name; null is Gymnasium's default, body 1 (the torso). Resolved at each step, as Gymnasium resolves it, so a bad one fails at the first step with MuJoCo's error.terminateWhenUnhealthyboolWhether an unhealthy state ends the episode (
terminate_when_unhealthy).healthyZRange(double hx, double hy)?The inclusive range of a healthy torso height (
healthy_z_range); null is (0.2, 1.0).contactForceRange(double hx, double hy)?The range the contact forces are clipped to (
contact_force_range); null is (-1.0, 1.0).resetNoiseScaledoubleThe scale of the reset's noise (
reset_noise_scale).excludeCurrentPositionsFromObservationboolWhether the root's x and y leave the observation (
exclude_current_positions_from_observation).includeCfrcExtInObservationboolWhether the clipped contact forces join the observation (
include_cfrc_ext_in_observation).widthintThe rendered frame's width (
width).heightintThe rendered frame's height (
height).cameraIdint?The camera to render from (
camera_id).cameraNamestringThe camera to render from by name (
camera_name).maxGeomintThe most geoms a frame draws (
max_geom).visualOptionsIReadOnlyDictionary<int, bool>MuJoCo visual flags to switch (
visual_options).viewerFactoryEnvViewerFactoryGym.NET's
viewer_factory(DR-20) forhumanmode.
Remarks
The mode is required and comes first, so new AntEnvV5() keeps meaning the other constructor, without a
mode; leave the mode out there for no rendering. Every other parameter means what it means there.
Exceptions
- ArgumentOutOfRangeException
Thrown for a value RenderMode does not declare (a cast integer such as
(RenderMode)99).- NotSupportedException
xmlFileis not a bundled model.- OverflowError
A frame skip of 0 (Python's
OverflowErrorderiving the frame rate).- ValueError
Both a camera id and a camera name.
AntEnvV5(string, int, IReadOnlyDictionary<string, object?>?, double, double, double, double, object?, bool, (double Low, double High)?, (double Low, double High)?, double, bool, bool, string?, int, int, int?, string?, int, IReadOnlyDictionary<int, bool>?, EnvViewerFactory?)
Creates the env (Gymnasium's AntEnv.init): the model, its spaces and the frame rate the time step gives.
public AntEnvV5(string xmlFile = "ant.xml", int frameSkip = 5, IReadOnlyDictionary<string, object?>? defaultCameraConfig = null, double forwardRewardWeight = 1, double ctrlCostWeight = 0.5, double contactCostWeight = 0.0005, double healthyReward = 1, object? mainBody = null, bool terminateWhenUnhealthy = true, (double Low, double High)? healthyZRange = null, (double Low, double High)? contactForceRange = null, double resetNoiseScale = 0.1, bool excludeCurrentPositionsFromObservation = true, bool includeCfrcExtInObservation = true, string? renderMode = null, int width = 480, int height = 480, int? cameraId = null, string? cameraName = null, int maxGeom = 1000, IReadOnlyDictionary<int, bool>? visualOptions = null, EnvViewerFactory? viewerFactory = null)
Parameters
xmlFilestringThe model: a bundled file name (
xml_file, DR-63).frameSkipintPhysics steps per env step (
frame_skip).defaultCameraConfigIReadOnlyDictionary<string, object>The free camera's settings (
default_camera_config); null is Gymnasium's default,{"distance": 4.0}, and an empty dictionary leaves MuJoCo's camera as it is.forwardRewardWeightdoubleThe weight of the forward velocity (
forward_reward_weight; Gymnasium's default is the int 1).ctrlCostWeightdoubleThe weight of the squared torques (
ctrl_cost_weight).contactCostWeightdoubleThe weight of the squared clipped contact forces (
contact_cost_weight).healthyRewarddoubleThe reward for each healthy step (
healthy_reward).mainBodyobjectThe body whose motion is the forward velocity (
main_body): its index (an integer, a bool or a 0-d integer array) or its name; null is Gymnasium's default, body 1 (the torso). Resolved at each step, as Gymnasium resolves it, so a bad one fails at the first step with MuJoCo's error.terminateWhenUnhealthyboolWhether an unhealthy state ends the episode (
terminate_when_unhealthy).healthyZRange(double hx, double hy)?The inclusive range of a healthy torso height (
healthy_z_range); null is (0.2, 1.0).contactForceRange(double hx, double hy)?The range the contact forces are clipped to (
contact_force_range); null is (-1.0, 1.0).resetNoiseScaledoubleThe scale of the reset's noise (
reset_noise_scale).excludeCurrentPositionsFromObservationboolWhether the root's x and y leave the observation (
exclude_current_positions_from_observation).includeCfrcExtInObservationboolWhether the clipped contact forces join the observation (
include_cfrc_ext_in_observation).renderModestringGymnasium's
render_mode:human,rgb_array,depth_array,rgbd_tupleor null.widthintThe rendered frame's width (
width).heightintThe rendered frame's height (
height).cameraIdint?The camera to render from (
camera_id).cameraNamestringThe camera to render from by name (
camera_name).maxGeomintThe most geoms a frame draws (
max_geom).visualOptionsIReadOnlyDictionary<int, bool>MuJoCo visual flags to switch (
visual_options).viewerFactoryEnvViewerFactoryGym.NET's
viewer_factory(DR-20) forhumanmode.
Exceptions
- NotSupportedException
xmlFileis not a bundled model.- OverflowError
A frame skip of 0 (Python's
OverflowErrorderiving the frame rate).- ValueError
Both a camera id and a camera name.
Properties
ClassMetadata
Gymnasium's class attribute metadata: the four render modes, the frame rate set per instance. A new dict on every read.
public static Metadata ClassMetadata { get; }
Property Value
ContactCost
The contact cost now (contact_cost): contact_cost_weight * sum(contact_forces**2).
[PyName("contact_cost")]
public double ContactCost { get; }
Property Value
ContactCostWeight
The weight of the squared clipped contact forces (_contact_cost_weight), read every step.
[PyName("_contact_cost_weight")]
public double ContactCostWeight { get; set; }
Property Value
ContactForceRange
The range the contact forces are clipped to (_contact_force_range), for the observation and the cost.
[PyName("_contact_force_range")]
public (double Low, double High) ContactForceRange { get; set; }
Property Value
ContactForces
The contact forces clipped to ContactForceRange (contact_forces), one row of 6 per body
(the world's included). A new (nbody, 6) array on every read, which the caller owns.
[PyName("contact_forces")]
public NDArray ContactForces { get; }
Property Value
CtrlCostWeight
The weight of the squared torques (_ctrl_cost_weight), read every step.
[PyName("_ctrl_cost_weight")]
public double CtrlCostWeight { get; set; }
Property Value
ExcludeCurrentPositionsFromObservation
Whether the root's x and y leave the observation (_exclude_current_positions_from_observation). Footgun:
changing it after construction changes the observation's length but not the observation space, as in Gymnasium.
[PyName("_exclude_current_positions_from_observation")]
public bool ExcludeCurrentPositionsFromObservation { get; set; }
Property Value
ForwardRewardWeight
The weight of the forward velocity (_forward_reward_weight), read every step.
[PyName("_forward_reward_weight")]
public double ForwardRewardWeight { get; set; }
Property Value
HealthyReward
The healthy reward now (healthy_reward): is_healthy * _healthy_reward, an np.float64 there.
[PyName("healthy_reward")]
public double HealthyReward { get; }
Property Value
HealthyRewardWeight
The reward for each healthy step (_healthy_reward), read every step.
[PyName("_healthy_reward")]
public double HealthyRewardWeight { get; set; }
Property Value
HealthyZRange
The inclusive range of a healthy torso height (_healthy_z_range).
[PyName("_healthy_z_range")]
public (double Low, double High) HealthyZRange { get; set; }
Property Value
IncludeCfrcExtInObservation
Whether the clipped contact forces join the observation (_include_cfrc_ext_in_observation).
[PyName("_include_cfrc_ext_in_observation")]
public bool IncludeCfrcExtInObservation { get; set; }
Property Value
IsHealthy
Whether the ant is healthy now (is_healthy): every position and velocity finite and the torso's height
inside the inclusive z range.
[PyName("is_healthy")]
public bool IsHealthy { get; }
Property Value
MainBody
The body whose motion is the forward velocity (_main_body): an index or a name, resolved at each step.
[PyName("_main_body")]
public object? MainBody { get; set; }
Property Value
ObservationStructure
How the observation is laid out (observation_structure): skipped_qpos, qpos, qvel and
cfrc_ext counts.
public Info ObservationStructure { get; }
Property Value
ResetNoiseScale
The scale of the reset's noise (_reset_noise_scale), read every reset.
[PyName("_reset_noise_scale")]
public double ResetNoiseScale { get; set; }
Property Value
TerminateWhenUnhealthy
Whether an unhealthy state ends the episode (_terminate_when_unhealthy), read every step.
[PyName("_terminate_when_unhealthy")]
public bool TerminateWhenUnhealthy { get; set; }
Property Value
Methods
ControlCost(NDArray)
The control cost of action (control_cost): ctrl_cost_weight * sum(action**2),
computed in the action's dtype and returned widened exactly.
[PyName("control_cost")]
public double ControlCost(NDArray action)
Parameters
actionNDArrayThe action.
Returns
- double
The cost.
Exceptions
- ArgumentNullException
actionis null.- NotSupportedException
An action of a dtype whose cost NumPy keeps in that dtype (float16, complex).
GetResetInfo()
The reset's info (_get_reset_info): x_position, y_position and
distance_from_origin of the root joint, np.float64s.
protected override Info GetResetInfo()
Returns
- Info
The info.
ResetModel()
The episode's start (reset_model): positions init_qpos + uniform(-scale, scale, nq), then
velocities init_qvel + scale * standard_normal(nv), drawn in that order.
protected override NDArray ResetModel()
Returns
- NDArray
The first observation.
Step(NDArray)
Applies action for FrameSkip physics steps and reports the transition (step).
public override StepResult<NDArray> Step(NDArray action)
Parameters
actionNDArrayThe 8 torques, float32 (the action space's samples) or any real dtype.
Returns
- StepResult<NDArray>
The observation, the float64 reward, whether an unhealthy state ended the episode,
truncated=Falseand the info.
Exceptions
- ArgumentNullException
actionis null.- ValueError
The action is not of shape
(8,), with Gymnasium's message.- NotSupportedException
An action of a dtype whose cost NumPy keeps in that dtype (float16, complex).
- KeyError
MainBody names no body (MuJoCo's message).
- IndexError
MainBody is an index outside the model's bodies (MuJoCo's message).
- TypeError
MainBody is neither an integer nor a string (pybind11's message).