rdabane

↗ HN profile [ 36.1 ms ] full profile
Karma
99
Created
March 24, 2008 (18y ago)
Submissions
0
  1. I'm looking for annotated training/test video data sequences for driver monitoring(Driver drowsiness etc) application development.

  2. Smallest size=memory required for intermediate + Network

  3. Assuming I've a MxN input image, I need to estimate minimum memory required for inference. I'm ok with any kind of reordering of computations.

  4. Why most of CNN papers use 224x224 or 223x223 size images? Why not use more standard like 640x480?