CoolFace
Datasetpublic

H1yori233/MultiCamVideo-Wan22

MultiCamVideo-Wan22 WanPRoPE preprocessing of KlingTeam/MultiCamVideo-Dataset for the Wan2.2 TI2V 5B backbone. Contents 136,000 rows from 13,600 synchronized scenes and 10 cameras per scene 32 logical shards, with scene groups kept intact 2,144 Parquet files: 67 files per shard 704×1280 center-cropped conditioning, 81 source frames at 15 fps 21 latent/action timesteps camera viewmats with shape [21, 4, 4] camera intrinsics Ks with shape [21, 3, 3]… See the full description on the dataset page: https://huggingface.co/datasets/H1yori233/MultiCamVideo-Wan22.

sourceHugging Faceapache-2.0updated 2mo agoView on Hugging Face
0likes86downloads
Dataset Card

MultiCamVideo-Wan22

WanPRoPE preprocessing of `KlingTeam/MultiCamVideo-Dataset` for the Wan2.2 TI2V 5B backbone.

Contents

  • —136,000 rows from 13,600 synchronized scenes and 10 cameras per scene
  • —32 logical shards, with scene groups kept intact
  • —2,144 Parquet files: 67 files per shard
  • —704×1280 center-cropped conditioning, 81 source frames at 15 fps
  • —21 latent/action timesteps
  • —camera viewmats with shape [21, 4, 4]
  • —camera intrinsics Ks with shape [21, 3, 3]

Preprocessing

  • —FastVideo source revision: 8cf07cb94fb5ac062ee430bd75df2e51cda1730a
  • —retention/EOF-flush fix: d21acfb6a4f21221ddd391fd28b07f4ab510e727
  • —model revision: b8fff7315c768468a5333511427288870b2e9635
  • —source dataset revision: 4e7d995480cfa915f7c3721e616ed14789d87ff7
  • —preprocessing batch size: 8
  • —samples_per_file=64
  • —flush_frequency=64

See metadata/ for the complete run manifest, validation report, resource snapshot, and source attribution.

License and attribution

The source MultiCamVideo dataset is licensed under Apache-2.0. This preprocessed release retains the same license and source attribution. See metadata/source-attribution.md.