storm_kit.mpc.cost package¶
-
class
DistCost
(weight=None, vec_weight=None, gaussian_params={}, device=device(type='cpu'), float_dtype=torch.float32, **kwargs)[source]¶ Bases:
torch.nn.modules.module.Module
Initializes internal Module state, shared by both nn.Module and ScriptModule.
-
_is_full_backward_hook
: Optional[bool]¶
-
forward
(disp_vec, dist_type='l2', beta=1.0, RETURN_GOAL_DIST=False)[source]¶ Defines the computation performed at every call.
Should be overridden by all subclasses.
Note
Although the recipe for forward pass needs to be defined within this function, one should call the
Module
instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.
-
training
: bool¶
-
-
class
FiniteDifferenceCost
(tensor_args={'device': device(type='cpu'), 'dtype': torch.float32}, weight=1.0, order=1, gaussian_params={}, **kwargs)[source]¶ Bases:
torch.nn.modules.module.Module
Initializes internal Module state, shared by both nn.Module and ScriptModule.
-
_is_full_backward_hook
: Optional[bool]¶
-
training
: bool¶
-
-
class
JacobianCost
(ndofs, device, float_dtype, retract_weight)[source]¶ Bases:
torch.nn.modules.module.Module
Initializes internal Module state, shared by both nn.Module and ScriptModule.
-
_is_full_backward_hook
: Optional[bool]¶
-
forward
(state_batch, ee_pos_batch, ee_rot_batch, goal_ee_pos, goal_ee_rot, jac_batch, dt, proj_type='transpose', dist_type='l2', weight=1.0, beta=1.0, retract_state=None)[source]¶ Defines the computation performed at every call.
Should be overridden by all subclasses.
Note
Although the recipe for forward pass needs to be defined within this function, one should call the
Module
instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.
-
training
: bool¶
-
-
class
PoseCost
(weight, vec_weight=[], position_gaussian_params={}, orientation_gaussian_params={}, tensor_args={'device': 'cpu', 'dtype': torch.float32}, hinge_val=100.0, convergence_val=[0.0, 0.0])[source]¶ Bases:
torch.nn.modules.module.Module
Rotation cost
\[\]r &= sum_{i=0}^{num_rows} (R^{i,:} - R_{g}^{i,:})^2 cost &= sum w dot r
Initializes internal Module state, shared by both nn.Module and ScriptModule.
-
_is_full_backward_hook
: Optional[bool]¶
-
forward
(ee_pos_batch, ee_rot_batch, ee_goal_pos, ee_goal_rot)[source]¶ Defines the computation performed at every call.
Should be overridden by all subclasses.
Note
Although the recipe for forward pass needs to be defined within this function, one should call the
Module
instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.
-
training
: bool¶
-
-
class
ProjectedDistCost
(ndofs, weight=None, vec_weight=None, gaussian_params={}, device=device(type='cpu'), float_dtype=torch.float32)[source]¶ Bases:
storm_kit.mpc.cost.dist_cost.DistCost
Initializes internal Module state, shared by both nn.Module and ScriptModule.
-
_is_full_backward_hook
: Optional[bool]¶
-
forward
(disp_vec, jac_batch, proj_type='transpose', dist_type='squared_l2', beta=1.0)[source]¶ Defines the computation performed at every call.
Should be overridden by all subclasses.
Note
Although the recipe for forward pass needs to be defined within this function, one should call the
Module
instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.
-
training
: bool¶
-
-
class
ZeroCost
(device=device(type='cpu'), float_dtype=torch.float64, hinge_val=100.0, weight=1.0, gaussian_params={}, max_vel=0.01)[source]¶ Bases:
torch.nn.modules.module.Module
Initializes internal Module state, shared by both nn.Module and ScriptModule.
-
_is_full_backward_hook
: Optional[bool]¶
-
forward
(vels, goal_dist)[source]¶ Defines the computation performed at every call.
Should be overridden by all subclasses.
Note
Although the recipe for forward pass needs to be defined within this function, one should call the
Module
instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.
-
training
: bool¶
-
Submodules¶
- storm_kit.mpc.cost.bound_cost module
- storm_kit.mpc.cost.capsule_collision_cost module
- storm_kit.mpc.cost.circle_collision_cost module
- storm_kit.mpc.cost.collision_cost module
- storm_kit.mpc.cost.cost_base module
- storm_kit.mpc.cost.dist_cost module
- storm_kit.mpc.cost.ee_vel_cost module
- storm_kit.mpc.cost.finite_difference_cost module
- storm_kit.mpc.cost.gaussian_projection module
- storm_kit.mpc.cost.image_collision_cost module
- storm_kit.mpc.cost.jacobian_cost module
- storm_kit.mpc.cost.manipulability_cost module
- storm_kit.mpc.cost.null_costs module
- storm_kit.mpc.cost.pose_cost module
- storm_kit.mpc.cost.primitive_collision_cost module
- storm_kit.mpc.cost.projected_dist_cost module
- storm_kit.mpc.cost.robot_self_collision_cost module
- storm_kit.mpc.cost.stop_cost module
- storm_kit.mpc.cost.voxel_collision_cost module
- storm_kit.mpc.cost.zero_cost module