Preprocess video data

Dear organizers,

Are we allowed to preprocess video data to sequence 2D image data before training ? If so, how many disk storage is available ?

Dear Fangyijie Wang,

Any preprocessing or postprocessing computations you do before training are allowed. Do note that this must be performed using your own computational resources.

You are allowed to preprocess and postprocess data within your container at inference time. You can set how much memory your algorithm gets assigned on your container settings pages, when you create an algorithm. A maximum of 32 GB is allowed. For any operations performed within your algorithm container at inference time (after submitting it to one of the two challenge phases), the maximum you set will be considered.

I hope this clarifies your questions! Sofía