panda-70m / README.md
multimodalart's picture
Update README.md
3080a22 verified
|
raw
history blame
428 Bytes
metadata
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
modalities: image

Panda 70M dataset by Snap Inc

70M video-caption pairs

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