---
task_categories:
- image-to-text
- text-to-image
- text-to-video
- image-to-video
language:
- en
pretty_name: Panda-70M
splits:
- name: train
- name: train_2m
- name: train_10m
- name: test
- name: validation

---

# [Panda 70M dataset by Snap Inc](https://snap-research.github.io/Panda-70M/)

### 70M video-caption pairs

Code for downloading: https://github.com/snap-research/Panda-70M/dataset_dataloading