MCPcopy Create free account

hub / github.com/abhaybd/GraspMolmo / functions

Functions161 in github.com/abhaybd/GraspMolmo

↓ 13 callersMethodget
(self, object_id: str, scan_id: int)
graspmolmo/eval/utils.py:52
↓ 5 callersMethodsample
(self, n_categories: int | None = None, replace=False)
graspmolmo/datagen/datagen_utils.py:147
↓ 4 callersFunctionimg_to_pc
(rgb: np.ndarray, depth: np.ndarray, cam_info: np.ndarray, mask: np.ndarray | None = None)
graspmolmo/eval/utils.py:19
↓ 4 callersFunctionmake_wall
(sidelength, center, normal, name)
graspmolmo/datagen/datagen.py:133
↓ 3 callersFunctioncvh
Get the convex hull of a mesh. Due to numerical issues, mesh.convex_hull is sometimes not watertight (despite by definition being watertight)
graspmolmo/subsample_grasps.py:19
↓ 3 callersMethodfrom_categories
(cls, data_dir: str, categories: list[str], load_kwargs: dict | None = None)
graspmolmo/datagen/datagen_utils.py:111
↓ 3 callersFunctionhomogenize
(arr: np.ndarray)
graspmolmo/datagen/datagen.py:89
↓ 3 callersFunctionload_unaligned_mesh_and_grasps
(data_dir: str, path: str)
graspmolmo/subsample_grasps.py:183
↓ 3 callersFunctionrejection_sample
(sampler_fn: Callable[[], U], condition_fn: Callable[[U], bool], max_iters: int = 1000)
graspmolmo/datagen/datagen_utils.py:34
↓ 3 callersFunctionsample_grasps
(grasps: list[np.ndarray], grasp_succs: list[np.ndarray], n_grasps: int)
graspmolmo/subsample_grasps.py:231
↓ 3 callersMethodto
(self, device: str)
graspmolmo/eval/mask_detection.py:28
↓ 3 callersMethodto
(self, device: str)
graspmolmo/eval/pointcloud.py:105
↓ 2 callersMethod_load_mesh
(self, category: str, obj_id: str, center: bool = True)
graspmolmo/datagen/datagen_utils.py:160
↓ 2 callersMethodcategories
(self)
graspmolmo/datagen/datagen_utils.py:141
↓ 2 callersFunctionconstruct_cam_K
(w: int, h: int, dfov: float)
graspmolmo/datagen/datagen_utils.py:74
↓ 2 callersFunctioncreate_plane
(width: float, depth: float, center: np.ndarray, normal: np.ndarray)
graspmolmo/datagen/datagen.py:95
↓ 2 callersFunctiondownload
(url: str, filename: str)
graspmolmo/eval/utils.py:10
↓ 2 callersFunctiondraw_grasp
Draws the grasp with lines in-place onto the image. Args: image: The image to draw on. cam_K: (3, 3) The camera intrinsic mat
graspmolmo/inference/utils.py:115
↓ 2 callersFunctionget_desc
(n_generated_obs: int)
graspmolmo/datagen/generate_obs.py:256
↓ 2 callersFunctionget_grasp_points
Args: pc: (N, 3) The point cloud of the scene. grasps: (K, 4, 4) The grasp poses to get the points for. Returns: The
graspmolmo/inference/utils.py:76
↓ 2 callersMethodgrasps
(self, category: str, obj_id: str)
graspmolmo/datagen/datagen_utils.py:173
↓ 2 callersFunctionis_zero_measure_2d
(p: Path2D)
graspmolmo/subsample_grasps.py:32
↓ 2 callersFunctionload_aligned_meshes_and_grasps
(data_dir: str, category: str, obj_ids: list[str], n_proc=16)
graspmolmo/subsample_grasps.py:193
↓ 2 callersFunctionlook_at_rot
(p1: np.ndarray, p2: np.ndarray)
graspmolmo/datagen/datagen_utils.py:65
↓ 2 callersFunctionprincipal_axis
(points)
graspmolmo/subsample_grasps.py:65
↓ 2 callersFunctionprocess
(pc: np.ndarray, n_neighbors: int)
graspmolmo/eval/pointcloud.py:16
↓ 2 callersFunctionretrieve
(args, client: OpenAI)
graspmolmo/synthetic_annotations.py:327
↓ 2 callersFunctionrot_distance
(rot_deltas: np.ndarray)
graspmolmo/subsample_grasps.py:150
↓ 1 callersMethod__init__
(self, use_rri=True, feature_model=None, nearest_neighbors=20)
graspmolmo/eval/pointcloud.py:46
↓ 1 callersMethod_pred
(self, image: Image.Image, task: str, verbosity: int = 0)
graspmolmo/inference/grasp_predictor.py:58
↓ 1 callersFunctionadjust_camera_pose
(cam_pose: np.ndarray, cam_K: np.ndarray, gripper_obb_verts: np.ndarray, width: int, height: int)
graspmolmo/synthetic_annotations.py:130
↓ 1 callersFunctionbackproject
Args: cam_K: camera intrinsic matrix (3, 3) depth: depth image (H, W) Returns: xyz: xyz coordinates of the points in
graspmolmo/datagen/generate_obs.py:93
↓ 1 callersFunctionblock_signals
(signals: list[int])
graspmolmo/synthetic_annotations.py:74
↓ 1 callersFunctionblock_signals
(signals: list[int])
graspmolmo/datagen/generate_obs.py:39
↓ 1 callersFunctionbuild_scene
(data: dict[str, any])
graspmolmo/datagen/generate_obs.py:61
↓ 1 callersFunctioncentroid_align
(A, B)
graspmolmo/subsample_grasps.py:53
↓ 1 callersFunctioncopy_assets
(args)
graspmolmo/preprocess_shapenet.py:33
↓ 1 callersFunctioncreate_collage
(views: list[Image.Image], nrows: int, ncols: int)
graspmolmo/synthetic_annotations.py:201
↓ 1 callersFunctioncreate_molmo_sample
(sample_id: str, image_path: str, task: str, grasp_pt: np.ndarray)
graspmolmo/package_tg_train.py:107
↓ 1 callersFunctioncreate_query
(object_category: str, object_id: str, grasp_id: int, collage: Image.Image)
graspmolmo/synthetic_annotations.py:229
↓ 1 callersFunctioncreate_query
(object_category: str, object_id: str, candidate_grasp: str, annotated_grasps: list[str])
graspmolmo/datagen/match_tasks_to_grasps_v2.py:117
↓ 1 callersFunctioncreate_scene
(object_mesh: trimesh.Trimesh, grasp: np.ndarray)
graspmolmo/synthetic_annotations.py:82
↓ 1 callersMethoddetect_bbox
(self, object_name: str, image: Image.Image)
graspmolmo/eval/mask_detection.py:34
↓ 1 callersMethoddetect_mask
(self, object_name: str, image: Image.Image)
graspmolmo/eval/mask_detection.py:53
↓ 1 callersFunctiondraw_grasp_points
Draws the grasp points in-place onto the image. Args: image: The image to draw on. cam_K: (3, 3) The camera intrinsic matrix.
graspmolmo/inference/utils.py:91
↓ 1 callersFunctionencode_image
(image: Image.Image)
graspmolmo/synthetic_annotations.py:224
↓ 1 callersFunctionestimate_normals
Args: xyz: xyz coordinates of the points in the camera frame (N, H, W, 3) Returns: normals: normals of the points in the came
graspmolmo/datagen/generate_obs.py:108
↓ 1 callersFunctionestimate_pca_rotation
(A, B)
graspmolmo/subsample_grasps.py:64
↓ 1 callersFunctionestimate_transform
Estimate optimal similarity transform (s, R, t) using Procrustes.
graspmolmo/subsample_grasps.py:99
↓ 1 callersFunctionfetch_annotations
(hydra_cfg: DictConfig)
graspmolmo/datagen/datagen.py:560
↓ 1 callersFunctiongenerate_floor_and_walls
(scene: ss.Scene, datagen_cfg: DatagenConfig)
graspmolmo/datagen/datagen.py:109
↓ 1 callersFunctiongenerate_lighting
(scene: ss.Scene, datagen_cfg: DatagenConfig)
graspmolmo/datagen/datagen.py:142
↓ 1 callersFunctiongenerate_renders
(renderer: pyrender.OffscreenRenderer, scene: pyrender.Scene)
graspmolmo/datagen/generate_obs.py:127
↓ 1 callersFunctiongenerate_scene
(datagen_cfg: DatagenConfig, annotations: list[Annotation], object_library: MeshLibrary, background_library: M
graspmolmo/datagen/datagen.py:474
↓ 1 callersFunctiongenerate_views
( renderer: pyrender.OffscreenRenderer, tr_scene: trimesh.Scene, n_views: int, cam_K: np.ndarr
graspmolmo/synthetic_annotations.py:148
↓ 1 callersFunctionget_annotated_grasp_library
Args: obs_dir: observations directory out_dir: output directory Returns: annotated_grasp_library: dictionary mapping
graspmolmo/datagen/match_tasks_to_grasps_v2.py:50
↓ 1 callersFunctionget_annotations_in_view
Returns: in_view_annots: list[Annotation] - N annotations in view in_view_grasps: np.ndarray - (N, 4, 4) grasps in scene frame
graspmolmo/datagen/datagen.py:441
↓ 1 callersFunctionget_args
()
graspmolmo/synthetic_annotations.py:52
↓ 1 callersFunctionget_args
()
graspmolmo/package_tg_train.py:22
↓ 1 callersFunctionget_args
()
graspmolmo/preprocess_shapenet.py:19
↓ 1 callersFunctionget_args
()
graspmolmo/eval/generate_taskgrasp_eval.py:14
↓ 1 callersFunctionget_args
()
graspmolmo/datagen/match_tasks_to_grasps_v2.py:23
↓ 1 callersFunctionget_args
()
graspmolmo/datagen/package_pointing_data.py:17
↓ 1 callersFunctionget_args
()
examples/visualize_tg_img.py:16
↓ 1 callersFunctionget_candidate_grasp_library
(tasks_spec_path: str, out_dir: str)
graspmolmo/datagen/match_tasks_to_grasps_v2.py:93
↓ 1 callersFunctionget_grasp_constraint_torch
Args: grasp: (K, 4, 4) The grasp pose. Returns: A: (K, 6, 3) and b: (K, 6) such that Ax <= b implies that x is in the grasp v
graspmolmo/package_tg_train.py:33
↓ 1 callersFunctionget_grasp_constraint_torch
Args: grasp: (K, 4, 4) The grasp pose. Returns: A: (K, 6, 3) and b: (K, 6) such that Ax <= b implies that x is in the grasp v
graspmolmo/inference/utils.py:26
↓ 1 callersFunctionget_grasp_points
Args: pc: (N, 3) The point cloud of the scene. grasps: (K, 4, 4) The grasp poses to get the points for. Returns: The
graspmolmo/package_tg_train.py:83
↓ 1 callersFunctionget_grasp_points_torch
Args: pc: (N, 3) The point cloud of the scene. grasps: (K, 4, 4) The grasp poses to get the points for. Returns: The
graspmolmo/package_tg_train.py:53
↓ 1 callersFunctionget_grasp_points_torch
Args: pc: (N, 3) The point cloud of the scene. grasps: (K, 4, 4) The grasp poses to get the points for. Returns: The
graspmolmo/inference/utils.py:46
↓ 1 callersFunctionget_matching_results
(openai: OpenAI, batch_id: str)
graspmolmo/datagen/match_tasks_to_grasps_v2.py:177
↓ 1 callersFunctionget_visible_points
Check if points are visible from the camera. "Visible" means that a ray from the camera to the point does not intersect the scene mesh before
graspmolmo/datagen/datagen.py:167
↓ 1 callersFunctiongrasp_dist
Distance between a grasp and a set of grasps.
graspmolmo/subsample_grasps.py:154
↓ 1 callersMethodgrasp_points
(self, category: str, obj_id: str)
graspmolmo/datagen/datagen_utils.py:180
↓ 1 callersFunctionicp_2d
(src_mesh: trimesh.Trimesh, target_mesh: trimesh.Trimesh, N=1000, max_iterations=50, tolerance=1e-5)
graspmolmo/subsample_grasps.py:35
↓ 1 callersFunctionin_frustum_annotations
Check if grasps are within the camera frustum. Args: datagen_cfg: DatagenConfig cam_K: (3, 3) camera intrinsic matrix
graspmolmo/datagen/datagen.py:191
↓ 1 callersFunctionkelvin_to_rgb
(kelvin: float)
graspmolmo/datagen/datagen_utils.py:43
↓ 1 callersFunctionlist_s3_files
(s3: S3Client, bucket_name: str, prefix: str)
graspmolmo/utils.py:15
↓ 1 callersFunctionmain
()
graspmolmo/synthetic_annotations.py:365
↓ 1 callersFunctionmain
()
graspmolmo/package_tg_train.py:167
↓ 1 callersFunctionmain
()
graspmolmo/preprocess_shapenet.py:181
↓ 1 callersFunctionmain
()
graspmolmo/subsample_grasps.py:285
↓ 1 callersFunctionmain
()
graspmolmo/eval/generate_taskgrasp_eval.py:21
↓ 1 callersFunctionmain
(cfg: DictConfig)
graspmolmo/datagen/generate_obs.py:260
↓ 1 callersFunctionmain
(cfg: DictConfig)
graspmolmo/datagen/split_data.py:9
↓ 1 callersFunctionmain
()
graspmolmo/datagen/match_tasks_to_grasps_v2.py:293
↓ 1 callersFunctionmain
()
graspmolmo/datagen/package_pointing_data.py:121
↓ 1 callersFunctionmain
(hydra_cfg: DictConfig)
graspmolmo/datagen/datagen.py:582
↓ 1 callersFunctionmain
()
examples/visualize_tg_img.py:63
↓ 1 callersFunctionnoncolliding_annotations
Check if grasps are noncolliding, i.e. the grasp is not in collision with any other object in the scene. Returns a mask of the same length as
graspmolmo/datagen/datagen.py:248
↓ 1 callersFunctionparse_point
Args: pred: The prediction string from the model. image_size: The size of the image, (width, height). If provided, return in pixe
graspmolmo/inference/grasp_predictor.py:13
↓ 1 callersFunctionpc_to_depth
(xyz: np.ndarray, cam_info: np.ndarray, height: int, width: int)
graspmolmo/eval/utils.py:31
↓ 1 callersFunctionpoint_on_support
(scene: ss.Scene)
graspmolmo/datagen/datagen.py:275
↓ 1 callersFunctionpoint_to_xml
(grasp_pt: np.ndarray)
graspmolmo/package_tg_train.py:99
↓ 1 callersFunctionpoint_to_xml
(grasp_pt: np.ndarray)
graspmolmo/datagen/package_pointing_data.py:54
↓ 1 callersFunctionpoints_to_tuples
(grasp_pts: np.ndarray)
graspmolmo/datagen/package_pointing_data.py:28
↓ 1 callersMethodpred_points
Args: image: The image of the scene. task: The task to predict the grasp point for. verbosity: The verbos
graspmolmo/inference/grasp_predictor.py:73
↓ 1 callersFunctionproject_grasps
(args)
graspmolmo/preprocess_shapenet.py:121
↓ 1 callersFunctionrandom_delta_rot
(roll_range: float, pitch_range: float, yaw_range: float)
graspmolmo/datagen/datagen_utils.py:83
↓ 1 callersMethodregister
template: (N, 3) source: (M, 3) returns: (4, 4) transformation matrix, (M, 3) transformed source, final cost
graspmolmo/eval/pointcloud.py:86
next →1–100 of 161, ranked by callers