Skip to content

Latest commit

 

History

History
15 lines (9 loc) · 566 Bytes

README.md

File metadata and controls

15 lines (9 loc) · 566 Bytes

Narration File Pre-processing

This section hosts the detail information to conduct narration file pre-processing for model pre-training.

First get official Egoclip file and the annotated training split files from visual/moment/natural language queries challenges.

python create_pretrain_data_from_egoclip.py
python filter_missing_video.py

The first code aims to generate the pretrained data, and the second code aims to filter some train samples.