I wanna build a model to read logs and classify them. How should I start? (preferably TensorFlow solutions). The input is gonna be log files.

  • 1
    $\begingroup$ Your question is lacking details IMO. In general you would include your own research so far in your question. 1.) To show you put in effort and 2.) to address you question with more accuracy. What did you find and what are your questions related to what you have found? On what will you classify? What is in the logs? Technical logs with compile errors? Logs of transactions? What problem are you trying to solve with the network? To try to help you a bit further. How many categories do you have? With 2 you could use binary_crossentropy, for more you could categorical_crossentropy from TF+Keras. $\endgroup$ – Mike de Klerk Mar 10 at 13:36
  • 2
    $\begingroup$ Hi and welcome to AI SE! Your current question is too broad, as it lacks details, as stated by @MikedeKlerk. Please, address the points in Mike's comment above, then I will re-open your post ;) $\endgroup$ – nbro Mar 10 at 20:47