- Add save_scene_graph() and load_scene_graph() for persisting scene graphs to JSON
- Implement object node overlay (colored by label) in render_topdown_scene_map
- Implement edge arrows (room → object) in render_topdown_scene_map
- Add TopDownRenderStyle fields for object/edge visual configuration
- Add scene graph caching to verification notebook to skip rebuilds
- Add render_scene_graph_birdseye cell for full scene graph visualization
- Add display_objects_by_room and display_room_summary cells
- Evaluate multiple datasets in a single run (CIFAR10 + CIFAR100)
- Report Recall@K for a list of K values from one underlying search
- Save results to disk: summary.md, metrics.csv, per-dataset predictions.csv, confusion_matrix.csv
- Richer evaluation output: query_labels, topk_ids, topk_labels for downstream analysis
- Add --dataset and --top-k CLI overrides for benchmark command
- Update config schema: dataset→datasets, top_k→top_k_list
- Add bbox_depth_center position strategy in SceneGraphBuilder using depth
at bbox centre and configurable camera_hfov_degrees for pinhole projection.
- Add optional depth_sensor_uuid to HabitatSimulatorConfig; create depth
sensor spec alongside RGB sensor.
- Add camera_position/camera_rotation fields to RoomView; capture pose from
sensor_states when depth sensor is available.
- Update flatten_room_views for backward compatibility with legacy tuple
format.
- Wired in depth sensor and bbox_depth_center strategy in verification
notebook.
- Add tests for depth sensor support and new position strategies.
Introduce SceneGraphBuilder + SceneGraphBuildConfig to decouple scene graph
construction from the verification notebook. The builder handles batch
inference, hash encoding, and object node creation internally.
- Add SceneGraphBuilder.build_from_room_views() as the main entry point
- Add SceneGraphBuildConfig for inference_batch_size and position strategy
- Add SceneGraphBuildArtifacts to carry cropped images and debug metadata
- Extend ObjectNode with optional detection metadata (label, confidence,
bbox_xyxy, source_view_id, position_confidence)
- Add RoomView frozen dataclass as a structured view container
- Add flatten_room_views() utility to replace inline list comprehensions
- Refactor notebooks/verification.py to use the new builder API
BREAKING CHANGE: ObjectNode now accepts additional optional fields; direct
scene_graph.objects[obj_id] = ObjectNode(...) construction in the notebook
is replaced by builder.build_from_room_views(...).
- Introduce ImageHashPipeline Protocol for extensible hash computation
- Rename query_crop_index to query_index for clarity
- Make query_crop nullable to handle missing crop edge case
- Add keyword-only arguments for better API clarity
- Handle empty scene graph objects gracefully
- Add comprehensive test coverage for query_image_against_scene_graph
- Export ImageSceneGraphQueryResult and query_image_against_scene_graph from scenegraph module
- Replace inline hamming-distance-based image matching with dedicated query_image_against_scene_graph function
- Improve top_matches structure by extracting similarity scores and hash_bytes from matches
- Add .codegraph/ to gitignore (machine-local data, should not be committed)
- Add CodeGraph configuration for multi-language indexing
- Add SoftwareCamIndex class with xnor_popcount_score for CAM-style matching
- Add CamMatch and SceneGraphMatch dataclasses for query results
- Add query_by_visual_hash method to SimpleSceneGraph
- Add comprehensive tests for SoftwareCamIndex and xnor_popcount_score
- Add write_training_metrics() in new compressors/training_metrics.py
for appending epoch/step/lr/component rows as JSON Lines
- Wire --metrics-path and --log-every CLI options into train.py, passing
them to the training loop so metrics rows are written every N steps
- Accept absolute metrics paths or paths relative to output directory
- Add quantization component to loss log alongside existing distill/contrastive
- Replace inline torch.device() with get_device() utility
- Add test_hash_training_metrics.py covering multi-row JSONL append
Infrastructure:
- Pin torch 2.7.1 + CUDA 12.8 index for Linux/Windows in pyproject.toml
- Add .justfile rsync upload recipe with .stignore exclusion
- Exclude **/__marimo__ from rsync in .stignore
Dependencies updated: numpy 2.4.5, pandas 3.0.3, black 26.5.0,
click 8.4.0, contourpy, etc.
- Extract model loading logic from benchmark CLI into task-owned prepare_benchmark
- Add RetrievalEncoder class wrapping DINO with optional hash compression
- Add accelerate dependency for device management
- Switch dataset from CIFAR-10 to CIFAR-100 with fine_label column
- Add FramePacket dataclass to encapsulate per-image pipeline state
- Rename internal methods with underscore prefix convention
- Replace separate filter_batch/crop_batch with unified process_batch method
- Update notebook to use new HashPipeline API
- Add Owlv2ForObjectDetection and Owlv2Processor imports to model_loader
- Refactor load_dino_model to return tuple of processor and model
- Rewrite generate_proposals_batch to group images by bbox count for efficient batching
- Add _normalize_single_bbox_list helper for bbox normalization
- Update verification.py to use new pipeline architecture with detect/segment/filter/crop steps
- Add crop_batch method to HashPipeline for cropping images using OWLv2 detection boxes
- Integrate crop_batch into pipeline forward pass (extract_hash and extract_features)
- Move extract_masked_region from compressors/proposal/utils.py to utils/image.py
- Add crop_image_by_bbox utility function in utils/image.py
- Update type annotations to use dict[str, Any] instead of bare dict
- Update .justfile to add memory server command
- Update marimo dependency to >=0.22.0
- Update nvidia CUDA package markers for platform compatibility
- Replace bbox-prompted segmentation with OWLv2 text-guided object detection
- Refactor HashPipeline from nn.Module to plain class with modular stage methods
- Add detect_batch, segment_batch, filter_batch for explicit pipeline stages
- Rename forward to forward_batch with text_labels API instead of bboxes
- Add mask_scoring_config, score_threshold, postprocess_threshold configuration
- Update model_loader to expose Dinov2Model type annotation
- Refactor model_loader.py: improve return type annotations from tuple[Any, Any] to tuple[AutoImageProcessor, AutoModel]
- Refactor proposal/core.py: add input validation for mask array dimensionality, handle 2D masks and batch dimensions gracefully
- Refactor proposal_segament.ipynb: replace inline model loading with centralized load_owlv2_model() and load_sam_model() functions, use batched detect_objects_batch() and generate_proposals_batch() APIs
- switch OWLv2 loader return type to Owlv2ForObjectDetection
- add detect_objects and detect_objects_batch with two-stage score filtering
- add DetectionResult typed dict and conversion helper for post-processed outputs
- export new detection APIs from proposal module
- Replace SAM2AutomaticMaskGenerator pipeline with Sam2Processor+Sam2Model
- Freeze SAM model parameters at load time, removing torch.no_grad() at call sites
- Rewrite proposal/core.py to use bbox prompts instead of automatic point sampling
- Add bboxes parameter to all HashPipeline public methods (forward, forward_dataset, extract_features, extract_features_dataset)
- Extract mask filtering logic (_filter_masks) from proposal into pipeline
- Rename object_score/ to filter/
- Add load_owlv2_model to model_loader
- Rename notebooks/test.py to habitat_sim_setup.py
- Remove dino_compressor.py and segament_compressor.py
- Rewrite pipeline.py to inline DINO into HashPipeline
- Maintain backward compatibility: SAMHashPipeline alias
- Update tests and benchmark.py
