Skip to content

README.md

Latest commit

 

History

History
43 lines (31 loc) · 1.36 KB

README.md

File metadata and controls

43 lines (31 loc) · 1.36 KB

Instruction for the related files in utils

Hugging Face

download_TSBOW

The download_TSBOW.py provides the python source code for download our TSBOW dataset from Hugging Face.

The download options are: videos, annotations, metadata, semilabels, comparison, all. You can select one of them to download the data type.

  • videos:      the .mp4 files
  • annotations:  the manually labeled images and annotations
  • metadata:   the json file storing metadata
  • semilabels:    the semi-labeled annotations (txt files)
  • comparison:  the manually labeled images and annotations for Datasets' Comparison in Experiments
  • all:       all files and folders

The download_TSBOW.py can be executed via command:

python download_TSBOW.py                        \
    --repo_id       'TSBOW'                     \
    --output_dir    'NGOCHDM/Dataset/TSBOW/'    \
    --type          'all'

or run the bash file:

bash download_TSBOW.sh

Library requirement: huggingface_hub

!pip install huggingface_hub

Notes: The first public version, we only provide 30% of the test set.