unreal.MLLevelSetModelAndBonesBinningInfo¶
- class unreal.MLLevelSetModelAndBonesBinningInfo(parent_bone_name: str = '', active_bone_names: str = '', ml_model_inference_info_data_table_path: str = '', ml_model_inference_info_data_table_index: str = '', ml_model_inference_for_incorrect_zone_info_data_table_index: str = '', number_of_rotation_components_per_bone: None = [], rotation_component_indexes: None = [], signed_distance_scaling: float = 0.0, debug_grid_resolution: None = [], reference_bone_rotations: None = [], reference_bone_translations: None = [], training_grid_origin: None = [], training_grid_axis_x: None = [], training_grid_axis_y: None = [], training_grid_axis_z: None = [])¶
Bases:
TableRowBaseMLLevel Set Model and Bones Binning Info
C++ Source:
Module: Engine
File: MLLevelSetModelAndBonesBinningInfo.h
Editor Properties: (see get_editor_property/set_editor_property)
active_bone_names(str): [Read-Write] The bones that are trained for deformation. We suggest only train one bone per MLLevelSet for efficiencydebug_grid_resolution(Array[int32]): [Read-Write] Resolution of the Grid that is used for debug visualizationml_model_inference_for_incorrect_zone_info_data_table_index(str): [Read-Write] The index of the DataTable that includes information about NNE Model for incorrect (safe-danger) zone. If uninitialized, this means no safe zone is used.ml_model_inference_info_data_table_index(str): [Read-Write] The index of the DataTable that includes information about NNE Model for signed distanceml_model_inference_info_data_table_path(str): [Read-Write] The path to the DataTable that includes information about NNE Modelnumber_of_rotation_components_per_bone(Array[int32]): [Read-Write] E.g. Use {1,2} if ActiveBone1 has one and ActiveBone2 has two active rotations.parent_bone_name(str): [Read-Write] The bone that MLLevelSet is attached to. Note that the deformations near this joint is not trainedreference_bone_rotations(Array[double]): [Read-Write] Reference Rotations for the Active Bones.reference_bone_translations(Array[double]): [Read-Write] Reference Translations for the Active Bones.rotation_component_indexes(Array[int32]): [Read-Write] The indices of the rotation components. E.g. use {1,1,2} if ActiveBone1 uses Rot.Y and ActiveBone2 uses Rot.Y and Rot.Z.signed_distance_scaling(double): [Read-Write] Usually MLModels for SDF are trained so that output lies in [-1,1]. To do so the SignedDistances (in the dataset) are divided by SignedDistanceScaling.SignedDistanceScaling is generally defined to be the max length of the training bounding box.
training_grid_axis_x(Array[float]): [Read-Write] First Edge of the bounding box that MLLevelSet is trained on.training_grid_axis_y(Array[float]): [Read-Write] Second Edge of the bounding box that MLLevelSet is trained on.training_grid_axis_z(Array[float]): [Read-Write] Third Edge of the bounding box that MLLevelSet is trained on.training_grid_origin(Array[float]): [Read-Write] Min Corner of bounding box that MLLevelSet is trained on.
- property active_bone_names: str¶
[Read-Write] The bones that are trained for deformation. We suggest only train one bone per MLLevelSet for efficiency
- Type:
(str)
- property debug_grid_resolution: None¶
[Read-Write] Resolution of the Grid that is used for debug visualization
- Type:
(Array[int32])
- property ml_model_inference_for_incorrect_zone_info_data_table_index: str¶
[Read-Write] The index of the DataTable that includes information about NNE Model for incorrect (safe-danger) zone. If uninitialized, this means no safe zone is used.
- Type:
(str)
- property ml_model_inference_info_data_table_index: str¶
[Read-Write] The index of the DataTable that includes information about NNE Model for signed distance
- Type:
(str)
- property ml_model_inference_info_data_table_path: str¶
[Read-Write] The path to the DataTable that includes information about NNE Model
- Type:
(str)
- property number_of_rotation_components_per_bone: None¶
[Read-Write] E.g. Use {1,2} if ActiveBone1 has one and ActiveBone2 has two active rotations.
- Type:
(Array[int32])
- property parent_bone_name: str¶
[Read-Write] The bone that MLLevelSet is attached to. Note that the deformations near this joint is not trained
- Type:
(str)
- property reference_bone_rotations: None¶
[Read-Write] Reference Rotations for the Active Bones.
- Type:
(Array[double])
- property reference_bone_translations: None¶
[Read-Write] Reference Translations for the Active Bones.
- Type:
(Array[double])
- property rotation_component_indexes: None¶
[Read-Write] The indices of the rotation components. E.g. use {1,1,2} if ActiveBone1 uses Rot.Y and ActiveBone2 uses Rot.Y and Rot.Z.
- Type:
(Array[int32])
- property signed_distance_scaling: float¶
[Read-Write] Usually MLModels for SDF are trained so that output lies in [-1,1]. To do so the SignedDistances (in the dataset) are divided by SignedDistanceScaling. SignedDistanceScaling is generally defined to be the max length of the training bounding box.
- Type:
(double)
- property training_grid_axis_x: None¶
[Read-Write] First Edge of the bounding box that MLLevelSet is trained on.
- property training_grid_axis_y: None¶
[Read-Write] Second Edge of the bounding box that MLLevelSet is trained on.