Sloguin's picture
|
download
raw
690 Bytes

TAO annotations

Files:

  • {train,validation}.json: Contains annotations for LVIS categories in TAO train and validation.
  • {train,validation}_with_freeform.json: Contains annotations for all categories in TAO train and validation.
  • test_without_annotations.json: JSON for test videos. The 'images' and 'videos' fields contain the images and videos that will be evaluated on the test set.
  • test_categories.json: Contains a list of categories which will be evaluated on the TAO test set.
  • video-lists/: Contains lists of videos and their source datasets contained in TAO. Please do not train on any videos in these lists! See video-lists/README.md for details.

Xet Storage Details

Size:
690 Bytes
·
Xet hash:
c0d67a9dee255555935788bf9a30f8882f753b08b22fe59809a5f1aa686af60d

Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.