1
$\begingroup$

Is there a golden rule which gives intuition on which base model needs to be used for a give image classification problem.

Most of the articles gives the below details which says how to train the model based on the dataset.

Rule for fine tuning the model (taken from analyticsvidya)

However I was not able to find good reference for the selection of base model.

Thank you

$\endgroup$
1
$\begingroup$

There is no specific rule associated with the base model selection for transfer learning. It is generally a trade-off between model precision and resource allocation. As the number of layers increases, the number of parameters increases and the model becomes more and more resource heavy but deeper models tend to have better accuracy over shallower counterparts. Here's a comparison:

conv arch

Apart from that also refer to this question: Which is the fastest image pre-trained model? if the speed of the model matters.

$\endgroup$

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.