My script to download the AVA2.1 dataset as of [Jun 1st, 2018].
- Download the annotation zipfile from [https://s3.amazonaws.com/ava-dataset/annotations/ava_v2.1.zip] as instructed in https://github.com/cvdfoundation/ava-dataset.
- Extract the content into the 'meta' directory.
- Execute the python script/submit job.
- Done!
A few minutes of your life is saved from replication efforts :).
Please consider using 'download_ava_threadpool.py' as a more automated script with multithreading.