---
dataset_info:
  features:
  - name: video_name
    dtype: string
  - name: video_bytes
    dtype: binary
  splits:
  - name: train
    num_bytes: 196938776843
    num_examples: 177454
  download_size: 196918851572
  dataset_size: 196938776843
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---

From [this video editing dataset](https://huggingface.co/datasets/QingyanBai/Ditto-1M) all source video, but in parquet