IKH

An Analysis of Deep Learning Models – 2

In the following lecture, we will continue our discussion on some other results related to inference time and ‘accuracy density’.

Let’s conclude the important points from the latter part of the paper:

  • Accuracy and inference time are in a hyperbolic relationship: a little increment in accuracy costs a lot of computational time.
  • Power consumption is independent of batch size and architecture.
  • The number of operations in a network model can effectively estimate inference time.
  • ENet is the best architecture in terms of parameters space utilization.

In the next segment, let’s summarise the topics that you learnt in this session.

Report an error