2
$\begingroup$

I recently developed a DNN model and I want to know what exactly is training time and what all steps are included in it? For ex I carried out the following steps 1) Determined best Network Architecture by 5 fold CV ( using Back-propagation) 2) Once I got the Best architecture I used unsupervised learning to get initial weights using unlabled data 3) I applied back-propagation to fine tune the network with initial weights given by step 2 (It ran for 500 epochs)

What is the training time in this procedure? Does it include all the steps or only steps 2 or 3? Or in general terms does the training time include time to run the code for the best architecture or it also includes the time taken to determine the best network architecture?

$\endgroup$
1
  • $\begingroup$ cpu time, sys time and total time..it completely depends on your models deepness.. $\endgroup$
    – Aditya
    Mar 22 '18 at 8:47
1
$\begingroup$

Generally, the most useful measure is the total "wall" time it takes to run the entire training script.

If training is defined as including hyperparameter search, then hyperparameters should be included. The result is the longest empirical time which is useful for estimating time it would take to reproduce similar results.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.