1
$\begingroup$

I do not know if i can ask this question here or not. But I really need it.

I'm very new to ML/DL/NN field.

I have seen many articles tackling the problem of selection of the best parameter for loop optimization like use ML/DL/NN to find best tile size only or best unrolling factor only or best vectorisation factor only.

So I wonder if is it possible to create a machine learning model that can at once find best parameter for loop unrolling, tiling and vectorisation together.

I got hard to reformulate my question but what I want to say whether is it possible to do a model that can learn for every loop optimization which is the best parameter. Will it require only to create a complexe model. To take more time to train it.

I tried to find an article to find a model in the internet that fix that but did not find yet. I asked this question cause i really have limited knowledge about machine learning.

Thank you to share any useful help or article that may answer my curiosity. Thank you again.

Honestly I have been afraid that my question will be misunderstood or taken as subjective, or get any pb so I hesitated at first but finally find if i do so will never learn so i get the courage to ask so I will accept every critic.

$\endgroup$
2
  • 2
    $\begingroup$ perhaps you're looking for "bayesian hyperparameter optimization"? $\endgroup$
    – shimao
    Jan 12, 2019 at 19:27
  • $\begingroup$ @shimao I'm null in ML and i do not know about bayesian hyperparameter optimization i have to search in the internet to understand it. Thank you very much $\endgroup$ Jan 13, 2019 at 16:09

0

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.