Functions — github.com/Lijiaxin0111/Open-d4rt

hub / github.com/Lijiaxin0111/Open-d4rt / functions

Functions626 in github.com/Lijiaxin0111/Open-d4rt

⨍ Functions 626◇ Types & classes 62

↓ 351 callersMethodget

(self, key: str)

src/core/registry.py:27

↓ 179 callersMethodget_path

(self, path: str, default: Any = None)

src/core/config.py:53

↓ 31 callersMethodkeys

(self)

src/core/registry.py:33

↓ 27 callersFunction_jsonable_float_array

(arr: np.ndarray, ndigits: int = 6)

vis/build_like_demo.py:41

↓ 14 callersMethod_tensor_sample

(self, value: Any, batch_index: int)

src/vis/training_visualizer.py:386

↓ 13 callersMethodmark_bad

( self, dataset: str, sample_key: str, sample_paths: list[str], failed

src/data/bad_sample_registry.py:147

↓ 11 callersFunction_np

(x: Any, dtype: np.dtype | None = None)

vis/serve_demo_viser.py:133

↓ 11 callersMethodis_bad_sample

(self, sample_key: str)

src/data/bad_sample_registry.py:129

↓ 11 callersFunctionis_retryable_data_error

Heuristic filter: skip only probable data I/O/corruption errors.

src/data/bad_sample_registry.py:37

↓ 11 callersFunctionmasked_mean

(values: torch.Tensor, mask: torch.Tensor, eps: float = 1e-6)

src/model/utils.py:31

↓ 10 callersFunction_bad_sample_max_retries

(cfg: Any)

src/data/builder.py:116

↓ 10 callersFunction_bad_sample_registry_path

(cfg: Any)

src/data/builder.py:112

↓ 10 callersFunction_clip_frames

(cfg: Any)

src/data/builder.py:108

↓ 10 callersFunction_hard_query_ratio

(cfg: Any)

src/data/builder.py:52

↓ 10 callersFunction_image_size

(cfg: Any)

src/data/builder.py:41

↓ 10 callersMethod_init_dataset_seeding

(self, namespace: str, default_seed: int = 0)

src/data/seeding.py:54

↓ 10 callersMethod_load_npy

(self, scene_dir: Path, name: str)

src/data/kubric_full_robust_preprocess_dataset.py:122

↓ 10 callersFunction_optional_int

(value: Any)

src/data/builder.py:120

↓ 10 callersMethod_prepare_sample_rng

(self, index: int, total: int, attempt: int)

src/data/seeding.py:92

↓ 10 callersFunction_prob_t_tgt_equals_t_cam

(cfg: Any)

src/data/builder.py:56

↓ 10 callersFunction_queries_per_clip

(cfg: Any)

src/data/builder.py:48

↓ 10 callersFunctionapply_photometric_augment

Apply temporally consistent color jitter/drop/blur to video in [0,1].

src/data/raw_augment.py:437

↓ 10 callersFunctionapply_spatial_crop_images_only

Random crop + resize-back on images/depth/K only (no query handling). Use this BEFORE query sampling so that queries are built in the cropped

src/data/raw_augment.py:614

↓ 10 callersFunctionaugment_cfg_from_train_config

(cfg: Any)

src/data/raw_augment.py:150

↓ 10 callersFunctionbuild_augment_info

Build a batch-stable augmentation info payload for debug/visualization. This must be present for every sample in a batch so that ``default_collat

src/data/raw_augment.py:85

↓ 10 callersFunctionfailed_paths_from_exception

(exc: Exception)

src/data/bad_sample_registry.py:50

↓ 10 callersMethodhas_any_bad_path

(self, paths: list[str])

src/data/bad_sample_registry.py:135

↓ 9 callersFunction_query_timestep_delta_kwargs

(cfg: Any)

src/data/builder.py:60

↓ 9 callersFunctionsample_frame_indices_with_stride

Select frame indices with optional temporal stride subsampling. When ``cfg.temporal_subsample_enabled`` is True and ``training`` is True, str

src/data/raw_augment.py:215

↓ 8 callersFunction_truthy

(value: Any)

src/data/raw_augment.py:114

↓ 8 callersMethodload_state_dict

(self, state: dict[str, Any])

src/engine/trainer.py:169

↓ 8 callersFunctionresolve_int

(value: Any, default: int)

src/model/utils.py:11

↓ 8 callersMethodset_path

(self, path: str, value: Any)

src/core/config.py:61

↓ 7 callersFunction_compute_average_pts_within_thresh

( gt_points: np.ndarray, pred_points: np.ndarray, scaling: str = "global", compute_epe: bool =

eval_track3d_in_worldtrack.py:231

↓ 7 callersFunction_dedup_str_list

(values: list[str])

src/data/kubric_full_robust_dataset.py:125

↓ 7 callersFunction_to_str_list

(value: Any)

src/data/builder.py:162

↓ 7 callersFunctionbuild_queries_from_depth

Build D4RT query/target/mask/query_stats from depth and camera. This assumes static-scene geometry for correspondence construction: a source-

src/data/depth_query_builder.py:88

↓ 7 callersFunctionsample_hard_query_flags

Sample per-query hard/easy flags with exact clip-level hard ratio.

src/data/raw_augment.py:316

↓ 6 callersFunction_encode_model_memory

( model: torch.nn.Module, video_b: torch.Tensor, aspect_b: torch.Tensor | None, )

src/eval/tasks.py:148

↓ 6 callersFunction_find_tensor

( src_state: dict[str, torch.Tensor], candidate_keys: tuple[str, ...] | list[str], expected_shape:

src/model/d4rt.py:71

↓ 6 callersFunction_run_model_for_queries

( model: torch.nn.Module, video_b: torch.Tensor, aspect_b: torch.Tensor | None, query: dict[st

src/eval/tasks.py:98

↓ 6 callersFunctionadd_handle

(h: Any)

vis/serve_demo_viser.py:320

↓ 6 callersMethodlog

(self, payload: dict[str, Any])

src/core/logging.py:39

↓ 6 callersMethodstate_dict

(self)

src/engine/trainer.py:166

↓ 5 callersFunction_as_np_scalar

(value: Any, dtype: Any)

src/data/kubric_full_robust_preprocess_dataset.py:55

↓ 5 callersMethod_failed_paths_from_exception

(self, exc: Exception, default: list[str] | None = None)

src/data/kubric_full_robust_dataset.py:312

↓ 5 callersMethod_model_module

(self)

src/engine/trainer.py:379

↓ 5 callersMethod_random_choice

(self, ids: np.ndarray, n: int, rng: np.random.Generator)

src/data/query_sampler.py:17

↓ 5 callersFunction_resize_nn

(arr: np.ndarray, out_hw: tuple[int, int])

src/data/kubric_full_robust_dataset.py:67

↓ 5 callersFunction_sort_value

(row: dict[str, Any], sort_key: str)

scripts/eval_worldtrack/batch_eval_worldtrack_step_ckpts.py:550

↓ 5 callersFunction_write_report

(report_path: Path, records: list[StepEvalRecord])

scripts/eval_worldtrack/batch_eval_worldtrack_step_ckpts.py:658

↓ 4 callersFunction_apply_crop_resize_extra

( arr_t_hw_or_hwc: np.ndarray, crop_xy: tuple[int, int], crop_hw: tuple[int, int], out_hw: tup

src/data/kubric_full_robust_dataset.py:74

↓ 4 callersFunction_as_int

(value: Any)

src/data_preprocess/prepare_kubric_full_preprocessed.py:121

↓ 4 callersFunction_build_query_for_targets

( query_uv_norm: np.ndarray, t_src: np.ndarray, t_tgt: np.ndarray, t_cam: np.ndarray, devi

infer_track_3d.py:219

↓ 4 callersFunction_estimate_sim3_closed_form

(src: np.ndarray, dst: np.ndarray)

eval_track3d_in_worldtrack.py:167

↓ 4 callersFunction_export_video_from_frames

(*, video_rgb: np.ndarray, fps: float, dst_video: Path)

vis/build_like_demo.py:1239

↓ 4 callersMethod_frame_indices

(self, scene_len: int, index: int)

src/data/scannet_raw_dataset.py:498

↓ 4 callersMethod_frame_paths

(self, scene: _Scene, frame_idx: int)

src/data/dynamic_replica_raw_dataset.py:557

↓ 4 callersFunction_model_clip_frames

(model: torch.nn.Module | None, default: int = 48)

src/eval/tasks.py:160

↓ 4 callersMethod_norm_path

(self, value: str | Path)

src/data/bad_sample_registry.py:71

↓ 4 callersFunction_normalize_supported_mode

(value: Any, *, path: str, supported: set[str])

src/data/raw_augment.py:130

↓ 4 callersMethod_reset_train_iter

(self)

src/data/kubric_full_robust_dataset.py:270

↓ 4 callersFunction_reshape

(pred: dict[str, torch.Tensor])

infer_track_3d.py:288

↓ 4 callersFunction_resolve_mixture_sources

(split: str, cfg: Any)

src/data/builder.py:583

↓ 4 callersFunction_run_point_cloud_queries_for_target_indices

( *, model: torch.nn.Module, video_clip: torch.Tensor, aspect_ratio: torch.Tensor | None,

vis/build_like_demo.py:772

↓ 4 callersMethod_save_fig_and_maybe_tb

(self, fig: Any, save_path: Path | None, tb_tag: str | None, step: int)

src/vis/training_visualizer.py:376

↓ 4 callersMethod_seed_material

(self, index: int, attempt: int, stream: int)

src/data/seeding.py:74

↓ 4 callersFunction_summary_json_complete

(path: Path)

scripts/eval_worldtrack/batch_eval_worldtrack_step_ckpts.py:131

↓ 4 callersFunction_tqdm_write

(progress: Any, message: str)

scripts/eval_worldtrack/batch_eval_worldtrack_step_ckpts.py:74

↓ 4 callersMethod_update_run_metadata

(self, update: dict[str, Any])

src/engine/trainer.py:571

↓ 4 callersMethod_validate

( self, loader: Any | None = None, max_batches_global: int | None = None, max_

src/engine/trainer.py:700

↓ 4 callersMethod_with_ignore_errors

(self, ds: Any)

src/data/kubric_full_robust_dataset.py:303

↓ 4 callersFunctionbuild_dataloader

( split: str, cfg: Any, manifest_arg: str | None = None, rank: int = 0, world_size: int =

src/data/builder.py:712

↓ 4 callersMethodclone

(self)

src/core/config.py:39

↓ 4 callersFunctiondepth_boundary_mask

Approximate hard-boundary mask from depth gradients.

src/data/raw_augment.py:278

↓ 4 callersFunctionload_checkpoint

(path: str | Path, map_location: str | torch.device = "cpu")

src/core/checkpoint.py:17

↓ 4 callersFunctionload_yaml_config

(path: str | Path)

src/core/config.py:71

↓ 4 callersFunctionsample_t_tgt_t_cam

Sample (t_tgt, t_cam) with exact clip-level P(t_tgt=t_cam) when feasible. When ``q_t_src`` and a delta-mixture are provided, ``t_tgt`` is sampled

src/data/raw_augment.py:386

↓ 4 callersMethodto_dict

(self)

src/core/config.py:42

↓ 3 callersMethod_checkpoint_payload

(self)

src/engine/trainer.py:382

↓ 3 callersMethod_clamp_t

(self, t: torch.Tensor)

src/model/query_embedding.py:77

↓ 3 callersFunction_compute_normal_map

Compute pseudo surface normals from a depth map via cross-product. Returns (normal_hw3, normal_valid_hw). Normals are in the camera coordinat

src/data/depth_query_builder.py:15

↓ 3 callersFunction_compute_point_motion_scores

( *, xyz_ref0: np.ndarray, visibility: np.ndarray, confidence: np.ndarray, )

vis/build_like_demo.py:195

↓ 3 callersFunction_dataset_children

(dataset: Any)

src/data/seeding.py:42

↓ 3 callersFunction_find_latest_event_file

(tensorboard_logdir: Path)

scripts/eval_worldtrack/batch_eval_worldtrack_step_ckpts.py:167

↓ 3 callersFunction_frame_id_from_stem

(stem: str)

src/data/blendermvs_raw_dataset.py:41

↓ 3 callersFunction_frame_id_from_stem

(stem: str)

src/data/mvs_synth_raw_dataset.py:43

↓ 3 callersFunction_make_anchor_clip_indices

(num_frames: int, clip_frames: int, target_idx: int, source_idx: int = 0)

infer_track_3d.py:89

↓ 3 callersFunction_meta_value

(meta: dict[str, Any], key: str, index: int)

src/vis/training_visualizer.py:70

↓ 3 callersFunction_nested_get

(payload: dict[str, Any], path: tuple[str, ...])

src/data_preprocess/prepare_kubric_full_preprocessed.py:85

↓ 3 callersFunction_next_non_comment_line

(handle)

src/data/blendermvs_raw_dataset.py:60

↓ 3 callersFunction_normalize_dataset_type

(dataset_type: str)

src/data/builder.py:129

↓ 3 callersFunction_pad_to_clip

(indices: list[int], *, target_len: int, valid_len: int)

src/data/raw_augment.py:236

↓ 3 callersMethod_poll_worldtrack_eval_processes

(self)

src/engine/trainer.py:405

↓ 3 callersMethod_reduce_validation_stats

( self, val_loss_sum: float, val_count: int, val_metric_sums: dict[str, float]

src/engine/trainer.py:796

↓ 3 callersFunction_resolve_tfds_dir

(input_root: Path)

src/data_preprocess/prepare_kubric_full_preprocessed.py:233

↓ 3 callersFunction_robust_vmax

(errors: np.ndarray, q: float = 0.95)

src/vis/training_visualizer.py:38

↓ 3 callersMethod_running_supervision_stats

(self)

src/engine/trainer.py:622

↓ 3 callersMethod_sample_rows

(pool: np.ndarray, count: int, rng: np.random.Generator)

src/data/kubric_full_robust_dataset.py:667

↓ 3 callersMethod_set_lr

(self, lr: float)

src/engine/trainer.py:146

next →1–100 of 626, ranked by callers