Step 1. Download the dataset from cvdfoundation/kinetics-dataset using the official download scrips.
bash download.sh k400_train_path.txt
# extract
bash extract.sh k400_train_path.txt
Step 2. Download new annotation (See the README.md from cvdfoundation/kinetics-dataset to get the latest URL)
wget https://s3.amazonaws.com/kinetics/400/annotations/train.csv
Step 3. Run make_subdirectory.py
to divide videos into subdirectories according to their label.
python make_subdirectory.py train.csv /path/to/video/folder
Just run k400.sh
, it will do all the things for you (unless you encounter some network error and the file is corrupted).
bash k400.sh