Santhosh Kolloju

@santhoshkolloju

NLP/NLU Artificial intelligence Engineer. My work involves applied research in area of Natural language processing for solving complex business problems.

Vrijeme pridruživanja: siječanj 2013.

Tweetovi

Blokirali ste korisnika/cu @santhoshkolloju

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

  1. proslijedio/la je Tweet

    Using Neural Networks for sales prospecting

    Poništi
  2. proslijedio/la je Tweet
    1. lis 2018.

    Here's How to Scrape 100 Google PAAs in 30s for Free 🤑 1. Download this free chrome plugin 2. Expand some PAAs 3. Right click on the title and scrape similar 4. Remove the [1] from Xpath and mash enter 5. Success

    Poništi
  3. proslijedio/la je Tweet
    4. velj

    We just released a Swedish BERT model, an alpha ALBERT-model and a BERT fine-tuned for NER based on texts from a wide range of sources (news, books, wikipedia, legal e-deposits, goverment publications, etc.).

    Prikaži ovu nit
    Poništi
  4. 4. velj
    Poništi
  5. proslijedio/la je Tweet
    4. velj
    Odgovor korisniku/ci

    If this is about , totally understand the sentiment & we are working hard on making sure the community can verify & reproduce. Our research (cc ) has always been very open whenever we can. Btw, conversations here are not cherry picked :)

    Poništi
  6. proslijedio/la je Tweet
    3. velj

    Check out BREAK - a new NLU benchmark for testing the ability of models to break down a question into the required steps for computing its answer. A work by Tomer Wolfson, accepted to TACL 2020.

    Poništi
  7. 2. velj

    ఆసక్తి అందరికి ఉంటుంది కాని ఆ శక్తి మాత్రం కొందరికె ఉంటుంది

    Poništi
  8. proslijedio/la je Tweet
    31. sij

    We are releasing HiPlot, a lightweight interactive visualization tool to help AI researchers discover correlations and patterns in high-dimensional data.

    Poništi
  9. proslijedio/la je Tweet

    High-perf, large transformer models across both CPU and GPU hardware. Microsoft developers use it to serve our customers. Learn how you can use to do it too:

    Poništi
  10. proslijedio/la je Tweet
    27. sij

    Reading list this week (Feb2) : 1. 's best practices to train nn's 2. Activation atlases by and 3. Christian perone's - MLE vs MAP 4. Reformers 5. 's/ 's latest vids. 6. 's news letter.

    Poništi
  11. proslijedio/la je Tweet
    25. sij

    BERT on steroids: Finetuning BERT for a dataset using pytorch and TPUs!

    Poništi
  12. proslijedio/la je Tweet
    25. sij

    Scaling Laws for Neural Language Models.

    Poništi
  13. proslijedio/la je Tweet
    23. sij

    Today gave a talk at Google Research Berlin about our Transformer analysis work. Met a lot of great people and feel totally happy :) Want to know details? Here are the blog posts: The Story of Heads: , Evolution of Representations:

    Poništi
  14. proslijedio/la je Tweet
    22. sij

    [Video Analysis] Reformer: The Efficient Transformer

    Poništi
  15. proslijedio/la je Tweet
    21. sij

    Nice summary of recent advances in language modeling.

    Poništi
  16. 21. sij
    Poništi
  17. proslijedio/la je Tweet
    18. sij

    Did you know that you can write blog posts for Pages using MS Word or Google Docs, including pasting images straight into your doc, and have it all synced to GitHub and included in your blog? Here's how it's done!

    Prikaži ovu nit
    Poništi
  18. 18. sij

    Guruji mass.. Director photo tho poster esaru ante.. 😎😎 Rajamouli tarvata trivikram maatrame.

    Poništi
  19. proslijedio/la je Tweet
    17. sij

    Thanks to *big* team effort, we released the code and the trained models from our LibriSpeech acoustic model architecture study and SOTA results here

    Poništi
  20. proslijedio/la je Tweet
    7. sij

    . people’s papers #2—ELECTRA: and colleagues (incl. at ) show how to build a much more compute/energy efficient discriminative pre-trainer for text encoding than BERT etc. using instead replaced token detection

    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:

    ·