results/omni_finetune/ANNOTATION_RECORD_PROBE.md

Xperience-10M Annotation Record Probe

Minimal-cost probe. Downloaded only annotation.hdf5; no MP4 or visualization.rrd files were downloaded.

• Repo: ropedia-ai/xperience-10m
• Probe count: 3
• Raw annotation cache: outside the published repo
• Local files only: False

9cecac72-8874-4b97-9541-18d4858f8e43/ep10/annotation.hdf5

• Downloaded annotation size: 6.38 MiB (6,687,192 bytes)
• HDF5 top-level keys: calibration, caption, depth, full_body_mocap, hand_mocap, imu, metadata, slam, video
• HDF5 dataset count: 65
• Largest first-dimension dataset: imu/accel_xyz with first dimension 190

Caption JSON Summary

Measure	Value
Parse status	ok
JSON bytes	1,178
Segment count	1
Current-action count	1
Object-frame count	1
Interaction-frame count	1
Sampled-frame count	1
Unique subtasks	1
Unique action labels	1
Unique objects	3
Action labels	["Arrange items in bin"]
Objects	["cardboard box", "hand", "plastic storage bin"]

Top Groups

Group	Dataset count	Max first dimension	First-dim histogram top values
calibration	23	4	{"4": 14}
caption	1	0	{}
depth	5	20	{"20": 2}
full_body_mocap	9	20	{"20": 9}
hand_mocap	10	20	{"20": 10}
imu	4	190	{"190": 3, "20": 1}
metadata	6	0	{}
slam	4	47	{"20": 3, "47": 1}
video	3	20	{"20": 2}

Caption / Action / Interaction Related Datasets

Dataset	Shape	Dtype	First dim	Sample values
caption	[]	object	None	["{"config": {"segment_sec": 20, "sample_fps": 0.5, "total_tokens": 2047, "Main Task": "Packing items into a plastic bin. The person is placing va...

cdc1ae12-a460-48ac-a892-7d314095c4b1/ep23/annotation.hdf5

• Downloaded annotation size: 6.38 MiB (6,687,256 bytes)
• HDF5 top-level keys: calibration, caption, depth, full_body_mocap, hand_mocap, imu, metadata, slam, video
• HDF5 dataset count: 65
• Largest first-dimension dataset: imu/accel_xyz with first dimension 188

Caption JSON Summary

Measure	Value
Parse status	ok
JSON bytes	1,051
Segment count	1
Current-action count	1
Object-frame count	1
Interaction-frame count	1
Sampled-frame count	1
Unique subtasks	1
Unique action labels	1
Unique objects	4
Action labels	["Pulling up sock"]
Objects	["bathroom floor", "feet", "sock", "toilet"]

Top Groups

Group	Dataset count	Max first dimension	First-dim histogram top values
calibration	23	4	{"4": 14}
caption	1	0	{}
depth	5	20	{"20": 2}
full_body_mocap	9	20	{"20": 9}
hand_mocap	10	20	{"20": 10}
imu	4	188	{"188": 3, "20": 1}
metadata	6	0	{}
slam	4	128	{"20": 3, "128": 1}
video	3	20	{"20": 2}

Caption / Action / Interaction Related Datasets

Dataset	Shape	Dtype	First dim	Sample values
caption	[]	object	None	["{"config": {"segment_sec": 20, "sample_fps": 0.5, "total_tokens": 2035, "Main Task": "Putting on socks. The person is standing in a bathroom and...

10282b64-a955-461e-9ef9-a1ddf8dc619a/ep5/annotation.hdf5

• Downloaded annotation size: 6.40 MiB (6,706,448 bytes)
• HDF5 top-level keys: calibration, caption, depth, full_body_mocap, hand_mocap, imu, metadata, slam, video
• HDF5 dataset count: 65
• Largest first-dimension dataset: slam/point_cloud with first dimension 837

Caption JSON Summary

Measure	Value
Parse status	ok
JSON bytes	1,299
Segment count	1
Current-action count	1
Object-frame count	1
Interaction-frame count	1
Sampled-frame count	1
Unique subtasks	1
Unique action labels	1
Unique objects	4
Action labels	["Walk down retail aisle"]
Objects	["person seated", "product packaging", "retail shelf", "shopping bags"]

Top Groups

Group	Dataset count	Max first dimension	First-dim histogram top values
calibration	23	4	{"4": 14}
caption	1	0	{}
depth	5	20	{"20": 2}
full_body_mocap	9	20	{"20": 9}
hand_mocap	10	20	{"20": 10}
imu	4	190	{"190": 3, "20": 1}
metadata	6	0	{}
slam	4	837	{"20": 3, "837": 1}
video	3	20	{"20": 2}

Caption / Action / Interaction Related Datasets

Dataset	Shape	Dtype	First dim	Sample values
caption	[]	object	None	["{"config": {"segment_sec": 20, "sample_fps": 0.5, "total_tokens": 2060, "Main Task": "walking through a retail store. The video shows a first-pe...