Download OVOW/engine/config/batch_config_example_legacy_rap.json from AIGeeksGroup/PhysicalWM: direct link, hf CLI and curl.
- Browser
- Download file 1.93 kB
-
https://huggingface.co/datasets/AIGeeksGroup/PhysicalWM/resolve/main/OVOW/engine/config/batch_config_example_legacy_rap.json
- Command line
-
hf download hf://datasets/AIGeeksGroup/PhysicalWM/OVOW/engine/config/batch_config_example_legacy_rap.json
-
curl -L -o batch_config_example_legacy_rap.json https://huggingface.co/datasets/AIGeeksGroup/PhysicalWM/resolve/main/OVOW/engine/config/batch_config_example_legacy_rap.json
1.93 kB
| { | |
| "//": "RAP 版本:使用 DA3 深度 + RAP 代替 FoundationPose,scale.mode=legacy 避免 DINO/FP 依赖", | |
| "input_root": "/mnt/einsia/aws01-nvme/einsia-shared/homes/gaomingju/jiayin/OVOW/data_pose/scenes", | |
| "scene_pattern": "scene_{:02d}", | |
| "metadata_filename": "scene.json", | |
| "keep_intermediate_video": false, | |
| "image_to_video": { | |
| "seconds": 7.0, | |
| "fps": 1.0, | |
| "min_side": 518, | |
| "codec": "libx264" | |
| }, | |
| "tracker": { | |
| "sample_stride": 3, | |
| "grid_size": 50, | |
| "vo_points": 756, | |
| "processing_mode": "da3", | |
| "da3_model_path": "depth-anything/DA3NESTED-GIANT-LARGE", | |
| "da3_process_res": 518, | |
| "da3_use_ray_pose": false, | |
| "da3_debug_depth": false, | |
| "single_frame_mode": false | |
| }, | |
| "generation": { | |
| "ss_guidance": 7.5, | |
| "ss_steps": 12, | |
| "slat_guidance": 3.0, | |
| "slat_steps": 12, | |
| "default_seed": 0, | |
| "randomize_seed": true, | |
| "depth_scene_step": 4, | |
| "scene_frame": -1, | |
| "export_scene_point_cloud": true, | |
| "object_mesh_samples": 15000, | |
| "model_type": "legacy" | |
| }, | |
| "scale": { | |
| "mode": "legacy" | |
| }, | |
| "rap": { | |
| "enable": true, | |
| "rap_root": "../RAP", | |
| "model": "rap_12", | |
| "flow_model_checkpoint": "weights/rap_model_12.ckpt", | |
| "feature_extraction_checkpoint": "weights/mini_spinnet_t.pth", | |
| "voxel_size": 0.25, | |
| "voxel_ratio": 0.05, | |
| "des_r": 5.0, | |
| "allocation_method": "voxel_adaptive", | |
| "min_points_per_part": 200, | |
| "max_points_per_part": 20000, | |
| "remove_outliers": true, | |
| "outlier_nb_neighbors": 20, | |
| "outlier_std_ratio": 2.5, | |
| "is_aligned_to_global_z": false, | |
| "use_random_downsample": false, | |
| "inference_sampling_steps": 10, | |
| "n_generations": 1, | |
| "rigidity_forcing": true, | |
| "generation": "generation_selected", | |
| "keep_rap_temp": false, | |
| "frame_stride": 1, | |
| "max_frames": -1, | |
| "point_stride": 2, | |
| "max_points_per_frame": 20000, | |
| "mesh_samples": 30000 | |
| } | |
| } | |