Rezultati pretraživanja
  1. 13. ruj 2018.

    May Google gift me one or two TPU's?

    Prikaži ovu nit
  2. prije 18 sati
    Odgovor korisniku/ci

    Really appreciate what you folks do!!

  3. 17. tra 2013.

    my classifier did worse than random... yeah!

  4. 4. lis 2016.

    "Athens" is currently trending, but the page includes tweets about Ohio, Georgia, and Greece.

  5. 2. velj
  6. 24. lip 2019.

    How I feel when I am tuning hyperparameters

  7. 25. stu 2015.

    Conversations with analytics: "That's not a bike, it's a seagull!" Tim

  8. Does anyone else feel sorry for the intern who had to aggregate everyone's data in ?

  9. Hmm, how will @Vitazdance do in explaining our Optimise personalisation engine next week to a non-techie audience?!

  10. 16. tra 2019.

    Research theme of the day: ‘It’s a random forest, not a magic forest’. Unfortunately.

  11. 6. lip 2013.

    Overfitting: Low error on Training data and high error on test data.

  12. 20. svi 2015.

    Training a classifier on the train ride home from JHH = 3rd degree thigh burns.

  13. my advice to kids: don't do drugs and study covariance matrix estimation.

  14. 1. svi 2019.
    Odgovor korisniku/ci

    Wrong way round, rocks are cheetahs and cows are elephants!

  15. 30. tra 2018.

    I know more about Bayes Ball than I do Baseball.

  16. 25. svi 2016.

    My Spotify recommendations were very fucked up after 10 years. New account, new data and it’s awesome now

  17. 14. ožu 2017.

    Don't know why a machine should recognize a corgi butt but still...

  18. 23. pro 2017.

    No matter how fast my computer is, it is always slow.

  19. 14. stu 2018.

    Now when an underground band gets popular not only do you have to listen to your uncool friend talk about it, his listening behavior in Spotify is now screwing up all your stations with the association matrix.

  20. 16. ruj 2014.

    Still annoyed that LDA stands for both "linear discriminant analysis" & "latent Dirichlet allocation".

Čini se da učitavanje traje već neko vrijeme.

Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.