Dataset Viewer
Auto-converted to Parquet Duplicate
Search is not available for this dataset
image
imagewidth (px)
832
1.54k
End of preview. Expand in Data Studio

LingBot Top2400-Compatible Training Data

Restricted internal training data. This dataset repository is private and is not intended for redistribution.

This is the self-contained training-data bundle referenced by the archived LingBot WoRL run. It contains the portable manifest together with every referenced first frame and camera-conditioning array.

Dataset summary

Item Value
Portable training records 2,218
First-frame images 2,218
Camera arrays 4,436
Total locked media assets 6,654
2-clip chains 912
3-clip chains 601
4-clip chains 705

The source collection contained 2,400 rows. During LingBot-compatible migration, 182 rows were excluded because their source stride did not match the required LingBot geometry. The exact excluded-row list is retained in skipped-source-rows.json.

Repository layout

manifest.jsonl              Portable training manifest used by WoRL
assets/first_frames/        First-frame conditioning images
assets/camera/              Camera poses and intrinsics in NumPy format
assets.lock.json            Size and SHA-256 lock for all referenced assets
migration-report.json       Migration identity and record counts
skipped-source-rows.json    Exact source rows excluded during migration
legacy-manifest.jsonl       Retained source-format manifest
dataset-card.md             Original restricted-data notice
EXPORT_RECEIPT.json         Export provenance and integrity summary

Each record in manifest.jsonl contains a stable ID and sampling index, a relative first-frame URI, relative camera pose and intrinsics URIs, camera metadata, and the ordered clip prompts/source windows. All paths are relative to the repository root.

Integrity

The WoRL dataset preflight completed successfully against the downloaded bundle:

records:             2218
semantic manifest:   d900daa6b760b6c7798e2840c550d80742e99043eb561d483628270bbcf08891
dataset revision:    b5cf7b1722628b208e33d13caa53d7c137686ab0769176ee5076cf0e9e835b81

assets.lock.json is authoritative for media-file sizes and SHA-256 digests. The semantic manifest digest is computed from the parsed canonical records; it is therefore intentionally different from the byte-level SHA-256 of manifest.jsonl, which is recorded in EXPORT_RECEIPT.json.

Access

The contents are marked INTERNAL-RESTRICTED. Keep the repository private and do not redistribute the data or referenced media.

Downloads last month
123