# Hugging Face — Build awesome datasets for video generation

- Company: Hugging Face (huggingface.co)
- Announced: 2025-02-12T00:00:00+00:00
- Category: developer-tool-release
- Subject: Platform
- Models affected: CogVideoX-5B, Qwen2VL
- Source: https://huggingface.co/blog/vid_ds_scripts
- Record: https://forck.live/items/1750-build-awesome-datasets-for-video-generation

Hugging Face released open source tooling and scripts for building video generation datasets, featuring a three-stage pipeline for acquisition, filtering, and processing. They demonstrated the tooling by creating datasets for 'crush' and 'cakeify' effects and fine-tuning the CogVideoX-5B model.

## Evidence

Verbatim from https://huggingface.co/blog/vid_ds_scripts:

> We have created various datasets with the tooling in an attempt to generate cool video effects, similar to the Pika Effects: Cakeify Crush We then used these datasets to fine-tune the CogVideoX-5B model using finetrainers.

---

Record: https://forck.live/items/1750-build-awesome-datasets-for-video-generation
Catalogue: https://forck.live/llms.txt
Feed: https://forck.live/feed.md