Dataset Viewer
Auto-converted to Parquet Duplicate
schema
string
collected_at
float64
refresh_seconds
int64
default_campaign
string
oakink_render
dict
dataset_release
string
execution_revision
string
queue_alive
bool
admission_held
bool
counts
dict
scope
dict
coverage
dict
throughput
dict
training_ready
null
known_gt_missing
list
auxiliary_fits
string
jobs
list
recent
list
hardware
dict
storage
dict
hardcases
dict
oakink_assets
dict
publisher
dict
history
list
egorenderer.public_progress.v1
1,794,161,453.897555
60
OakInk2
{ "dataset": "OakInk2", "package_index": "OakInk2/packages/final_native289_v354_20261102/INDEX.json", "execution_revision": "378", "queue_alive": true, "admission_held": true, "counts": { "complete": 567, "running": 3, "ready": 52, "failed": 0, "waiting_assets": 0, "total": 622 }, ...
V006
353
false
false
{ "complete": 301, "running": 0, "ready": 0, "failed": 0, "total": 301 }
{ "sources": 301, "source_frames": 218273, "variants": 1, "expected_variant_frames": 218273, "historical_extra_complete": 2, "cancelled_extra": 600 }
{ "completed_variant_frames": 218273, "sources_with_any_variant": 301, "sources_with_all_variants": 301 }
{ "packages_last_hour": 0, "queued_eta_hours": null }
null
[ { "source": "arctic_s01_capsulemachine_grab_01", "variant": 1 }, { "source": "arctic_s09_laptop_use_03", "variant": 1 } ]
PENDING_UPSTREAM_HANDOFF
[]
[ { "at": 1793599829.9072523, "source": "arctic_s10_laptop_use_02", "variant": 1 }, { "at": 1793599581.3385708, "source": "arctic_s10_notebook_use_02", "variant": 1 }, { "at": 1793599567.6335635, "source": "arctic_s10_waffleiron_use_02", "variant": 1 }, { "at": 1793...
{ "gpus": [ { "index": 0, "utilization": 0, "used_mib": 554, "total_mib": 97887 }, { "index": 1, "utilization": 88, "used_mib": 5731, "total_mib": 97887 }, { "index": 2, "utilization": 0, "used_mib": 5, "total_mib": 97887 ...
{ "ssd_available_gib": 698.6, "nas_available_gib": 2736, "ssd_budget_gib": 896, "ssd_reserved_gib": 415.5, "nas_active": 2, "nas_limit": 4 }
{ "complete": 15, "expected": 15, "frames": 10838, "training_ready": false }
{ "counts": { "COMPLETE": 622 }, "total": 622, "final_render_count": null }
{ "host": "PRO6000", "independent_of_local_codex": true, "refresh_seconds": 60 }
[ { "at": 1794154253.8799133, "complete": 301, "oakink_complete": 564 }, { "at": 1794154313.8814006, "complete": 301, "oakink_complete": 564 }, { "at": 1794154373.8846788, "complete": 301, "oakink_complete": 564 }, { "at": 1794154433.8837175, "complete": 301, ...

YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

EgoRenderer Public Progress

Read-only, sanitized status snapshots for the dataset research Space. This is telemetry, not a training dataset or a data download.

status.json uses egorenderer.public_progress.v1. collected_at is Unix time in seconds. The page polls every60seconds and marks data stale after180seconds. Counters distinguish original sources, randomized packages, rendered frames, known missing GT, and auxiliary-fit readiness. An archived or rendered package is not automatically a training-approved package.

No server credentials, host IPs, PIDs, filesystem paths, raw logs, original videos or model weights are published. The collector is read-only. Its finite local publishing process can go offline; stale snapshots must not be interpreted as live queue state.

Downloads last month
10,825