stereo-depth
zwm-stereo-depth-eval
ZWM stereo-depth eval dataset
Stereo image pairs with two annotated points per pair (with ground-truth
depth) used to evaluate depth perception in the ZWM repo.
The eval reuses ZWM's counterfactual flow engine: predict EPE for each
point's "motion" between stereo halves, then order depths by EPE magnitude
(closer = larger EPE).
Layout
stereo_depth/
├── dataset.json # 328 entries (164 pairs across kitti_1000 + kitti_500_flipud)
└── frames/
├──… See the full description on the dataset page: https://huggingface.co/datasets/awwkl/zwm-stereo-depth-eval.so101_stereo_depth_51UW-StereoDepth-40KStereo_Depth
Youtube Self Depth QA
Release status: private review only. Do not make this repository public until every source video/image has a documented redistribution right and the Level-3 validator is independently completed.
Dataset summary
Youtube Self Depth QA is a visual spatial-reasoning benchmark generated from relative disparity maps. It has three task levels:
Level 1: identify nearest/farthest marked point, optionally relative to a reference point.
Level 2:… See the full description on the dataset page: https://huggingface.co/datasets/LLDDSS/Stereo_Depth.so101_stereo_depth_This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "so101_follower",
"total_episodes": 3,
"total_frames": 673,
"total_tasks": 1,
"total_videos": 6,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 30,
"splits": {
"train": "0:3"
},
"data_path": "data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/aurora-skyfish/so101_stereo_depth_.so101_stereo_depth_54This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.1",
"robot_type": "so101_follower",
"total_episodes": 40,
"total_frames": 5186,
"total_tasks": 1,
"total_videos": 80,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 30,
"splits": {
"train": "0:40"
},
"data_path": "data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/aurora-skyfish/so101_stereo_depth_54.