Download and create a tfreader for the audioset dataset
cd download/train
cat ../balanced_train_segments.csv | ../download.sh
cd ../validate
cat ../eval_segments.csv | ../download.sh
cd download
python audioset_writer.py train
python audioset_writer.py validate
from audioset_generator import get_batchs
# get_batchs is a generator
my_gen = get_batchs(batch_size=10, n_epochs=1, num_threads=4)
my_gen.next()