0
$\begingroup$

I have built a dataset that I would like to pass to a pretrained model in oder to perform some predictions. I am looking for some steps/processes to guide me in this. Should I fine tune?If so what exactly should I fine tune? Is there other ways?

Any help is appreciated :)

$\endgroup$
1
  • $\begingroup$ I am new here too, but I would suggest reading blogs.rstudio.com/ai/posts/… if you are using Keras with R. It's about using pre-trained models and different types of fine-tuning. $\endgroup$
    – user122719
    Aug 7, 2021 at 14:05

1 Answer 1

0
$\begingroup$

If there is high similarity in your dataset and pretained model dataset and

  • If you have large amount of training samples, fine tune all layers of neural network model
  • If you have small number of training samples,fine tune last few layers of model

If similarity between your dataset and pretrained model dataset is low and

  • If you have large number of training samples, fine tune all layers or train the model from scratch.
  • If you have small number of training samples, it is difficult to get good model performance. you can select a less complex network and train it with heavily augmented data or acquire more data.
$\endgroup$

Your Answer

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

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