Rachael Tatman

@rctatman

Developer Advocate 🥑 . Linguistics PhD. Data science, NLP, Stats, ML, R, Python. Tweets my own & CC-BY-SA. (The 🦔's name is Gustav.) She/her

Vrijeme pridruživanja: siječanj 2010.

Tweetovi

Blokirali ste korisnika/cu @rctatman

Jeste li sigurni da želite vidjeti te tweetove? Time nećete deblokirati korisnika/cu @rctatman

  1. Prikvačeni tweet
    28. sij

    New live-streaming schedule my friends! I'm going back to twice a week @ 9:00 AM Pacific/5:00 PM GMT/10:30 PM IST. ⌚️ Wednesdays: Livecoding for now, might do other stuff in the future 🤓 ⌚️ Fridays: Livecoding! 💻 Tune in: ▶️ ▶️

    Poništi
  2. prije 18 sati

    Today I saw a bike with one of those trailers for bringing children with you. Dear readers, it was full of some very happy ferrets. 🚲 :ferret:

    Poništi
  3. 1. velj

    Bathtime for Gustav! 🦔✨

    Poništi
  4. proslijedio/la je Tweet

    I tell people to stop worrying about futuristic runaway smart AI, because it's stupid current AI that will kill us, this is what I mean. Someone bought a product that *literally* text-matched the word "charity" as 'good' and "vodka" as 'bad' and is making hiring choices with it.

    Prikaži ovu nit
    Poništi
  5. 1. velj

    Been puttering along most of the day with stuff with mixed success. Finally I just gave up and uninstalled and reinstalled the library that was giving me problems and... 🥁 🥁 🥁 It works fine now. Should have tried that three hours ago. 😅

    Poništi
  6. proslijedio/la je Tweet
    31. sij
    Poništi
  7. 31. sij

    Starting now! :)

    Prikaži ovu nit
    Poništi
  8. 30. sij

    Tomorrow, we'll be building a form and training our assistant for the first time! (Assuming everything goes to plan, ofc....🙃) Either way we should have a lot of fun. Come hang out and code with me. 🤓☕️

    Prikaži ovu nit
    Poništi
  9. 30. sij

    I've never done dev work on Windows before and woooow it shows. It took me like an hour just to get Docker running. 😅 Shout out to folks from the stream this morning who let me know that sudo doesn't work on Windows & vscode for automatically switching all my slashes. 🙏😂

    Poništi
  10. 29. sij

    Starting now!

    Prikaži ovu nit
    Poništi
  11. 28. sij

    Tomorrow morning we'll be back to working on our Dialect Quiz bot. No spoilers, but I've been working on it a bit on my own and I've... some good news and some bad news. 👀😳 Come hang out with me tomorrow! 💻🤓☕️

    Prikaži ovu nit
    Poništi
  12. 26. sij

    Gustav was Extremely Grumpy after his bath this week and sulked over to pout in one of his pipes. 😡🦔😂 (You can juuust see his little feet sticking out under his tummy.)

    Poništi
  13. 24. sij

    the macbook pro fan when you open a second chrome tab

    Poništi
  14. 23. sij

    I'm back, y'all! 💪🤓 (This week will probably be more project scoping than actual coding but we may tackle... installing things. 🙀)

    Poništi
  15. 23. sij

    (No promises if you what to know something genuinely super hard like "what are words?" 😱😰)

    Prikaži ovu nit
    Poništi
  16. 23. sij

    ML/Software folks: What language-y or NLP topics would you really appreciate a practical intro to? 🤔❓

    Prikaži ovu nit
    Poništi
  17. 23. sij

    Playing around with some slow motion video of Gustav. You can really see his little nose going! 🥰🦔👃

    Poništi
  18. 22. sij

    If you'd prefer watching to reading, the video is on 's YouTube channel. 🤓

    Prikaži ovu nit
    Poništi
  19. 22. sij

    Interesting finding from masked language models (i.e. BERT) are much better at preserving token identity than either machine translation or traditional language models. (Joint work w/ & . More info here: )

    A figure comparing t-SNE representations of is, are, were, was across the layers of machine translation, language model and masked language model (deeper on the right, shallower on the left). Only the masked language model maintains separate representations across all layers: language models collapses them to one blob and machine translation to two(ish).
    Prikaži ovu nit
    Poništi
  20. proslijedio/la je Tweet
    20. sij

    Thread with thoughts on : It is worth setting the work we do at in its historical context and think about future perspectives:

    Prikaži ovu nit
    Poništi

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.

    Možda bi vam se svidjelo i ovo:

    ·