Ivan Bilan

@DemiourgosUA

Data Scientist & Data Engineer for Natural Language Processing. All views are my own.

Munich
Vrijeme pridruživanja: srpanj 2014.

Tweetovi

Blokirali ste korisnika/cu @DemiourgosUA

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

  1. Prikvačeni tweet
    1. srp 2019.

    A year ago when I gave a presentation explaining the Transformer model, there were only a handful of published papers on the topic. Look at how far we've got, BERT, GPT-2, and now XLNet () are all contributing to the new leap in NLP.

    Poništi
  2. prije 8 sati

    GitHub Repo Spotlight №19 Today's pick is a library that can extract structured information from any kind of sentence: Snips is a great library for NLU. It can transform text into a structured JSON format.

    Poništi
  3. 4. velj

    GitHub Repo Spotlight №18 Today's pick is a library for scalable NLP, called Spark-NLP: It runs on Spark, either on a cluster of CPUs or GPUs. It is also easily configurable for both in-house cluster setups and cloud-based ones.

    Poništi
  4. 3. velj

    GitHub Repo Spotlight №17 a deep learning library that can be used to train Transformer models (like BERT, ALBERT, DistilROBERTA and more) quickly and effectively, it's called fast-bert:

    Poništi
  5. 2. velj

    GitHub Repo Spotlight №16 Today's pick is a package that merges the functionalities of one of the best machine learning libraries, scikit-learn, with one of the arguably best deep learning libraries, PyTorch. Its name is skorch:

    Poništi
  6. 1. velj

    GitHub Repo Spotlight №15 With FitBERT you can fill in the blanks within a sentence, namely you can mask out any word in the sentence and given a list of replacement options for that word, FitBERT will select the best one:

    Poništi
  7. 31. sij

    GitHub Repo Spotlight №14 An NLP toolkit that is built around sentence understanding tasks, its called jiant: Jiant will help you fast and effectively pre-train transfer learning models for various multitask learning problems.

    Poništi
  8. 30. sij

    GitHub Repo Spotlight №13 A closed domain question answering system called cdQA: cdQA uses BERT to create a question answering system for various specific domains and it would be a great addition to improve your Intranet search.

    Poništi
  9. 29. sij

    GitHub Repo Spotlight №12 A package that incorporates almost all imaginable readability index functions, it's called "textstat": With "textstat" you can measure how complex any given kind of text is.

    Poništi
  10. 28. sij

    GitHub Repo Spotlight №11 An implementation of a Plug and Play Language Models (PPLM) from Uber: If you work a lot with text generation and you are having problems running GPT-2 in production, your best bet is to try PPLM.

    Poništi
  11. 27. sij

    GitHub Repo Spotlight №10 Today's pick is a package for cleaning tabular data, PyJanitor: It gives you access to a lot of cleaning functions that can make your DataFrames more consistent.

    Poništi
  12. 26. sij

    GitHub Repo Spotlight №9 Today's pick is a package that let's you automatically extend your textual training data and generate similar sentence examples using Transformer based models:

    Poništi
  13. 25. sij

    - GitHub Repo Spotlight №8 Today's pick is a great library that enables data scientists and data engineers to write data related tests faster. It's called "great expectations":

    Poništi
  14. proslijedio/la je Tweet
    24. sij

    is now using adjacent sentences as context! I mentioned on the TAUS call that I had noticed DeepL seems to be doing it. Just now showed me an example that confirms it. Beautiful to see this is now reality.

    Prikaži ovu nit
    Poništi
  15. proslijedio/la je Tweet
    24. sij
    Poništi
  16. 24. sij

    - GitHub Repo Spotlight №7 Today's pick is a compendium of all latest impactful NLP papers from the top NLP conferences: It currently comprises links to papers from ACL and EMNLP for several years.

    Poništi
  17. 23. sij

    - GitHub Repo Spotlight №6 NLP library that incorporates many Deep Learning-based models into one easy to use package called gobbli:

    Poništi
  18. 22. sij

    - GitHub Repo Spotlight №5 Seq2seq library Headliner: , a great way to train and deploy your seq2seq models.

    Poništi
  19. 21. sij

    - GitHub Repo Spotlight №4 Today's pick is an Entity Matching approach that allows you to pre-train a Deep Learning model on any labeled data you might have:

    Poništi
  20. proslijedio/la je Tweet
    19. sij

    ⏲️ As of today, we have about eighteen years to go until the Y2038 problem occurs. But the Y2038 problem will be giving us headaches long, long before 2038 arrives. I'd like to tell you a story about this.

    Prikaži ovu nit
    Poništi
  21. 20. sij

    As usual, the rest of the list can be found under:

    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:

    ·