Thomas Wolf

@Thom_Wolf

Natural Language Processing, Deep Learning and Computational Linguistics - I lead the science team 🤗 He/him

Vrijeme pridruživanja: veljača 2011.

Tweetovi

Blokirali ste korisnika/cu @Thom_Wolf

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

  1. proslijedio/la je Tweet
    prije 6 sati

    104: and talk to us about model distillation, when you try to approximate a large model's decision boundary with a smaller model. After talking about the general area, we dive into DistilBERT.

    Poništi
  2. proslijedio/la je Tweet
    prije 10 sati

    The 2.4.0 release of transformers is **𝐌𝐀𝐒𝐒𝐈𝐕𝐄** thanks to our amazing community of contributors. 🔥

    Prikaži ovu nit
    Poništi
  3. proslijedio/la je Tweet
    2. velj

    I'm impressed by the work Hugging Face is doing.

    Poništi
  4. proslijedio/la je Tweet
    31. sij

    Turkish-: Anyone interested in a Turkish BERT and wants to evaluate it on downstream tasks? I did evaluation only for UD PoS tagging - any help is really appreciated! Would really like to have a proper evaluation before adding it to the Transformers hub🤗

    Prikaži ovu nit
    Poništi
  5. proslijedio/la je Tweet
    31. sij

    Supervised multimodal bitransformers now available in the awesome HuggingFace Transformers library!

    Poništi
  6. proslijedio/la je Tweet
    31. sij

    Transformers 2.4.0 is out 🤗 - Training transformers from scratch is now supported - New models, including *FlauBERT*, Dutch BERT, *UmBERTo* - Revamped documentation - First multi-modal model, MMBT from , text & images Bye bye Python 2 🙃

    Prikaži ovu nit
    Poništi
  7. proslijedio/la je Tweet
    31. sij

    Learn to build an interactive Transformer attention visualization based on and in under 30 minutes! We developed a minimal teaching example for our IAP class, publicly available here:

    Prikaži ovu nit
    Poništi
  8. 30. sij

    People asking me to teach classes clearly give zero fuck to the imposter syndrome of a former physics PhD turned lawyer before joining AI Anyway I'll co-teach NLPL Winter School w Yoav Goldberg talking transfer learning, its limits & where the field might head Will share slides

    Poništi
  9. 21. sij

    Let me highlight this amazing work I've read recently on in NLP, in which you'll find both: - a deep discussion of what it means for a neural model to be compositional - a deep and insightful comparison of LSTM, ConvNet & Transformers! 👉

    Poništi
  10. 19. sij
    Poništi
  11. 15. sij

    That’s a neat use of Transfer Learning to leverage a pretrained GAN (here BigGAN)

    Poništi
  12. 14. sij

    => you need to keep everything clear & visible. No unnecessary user-facing abstractions or layers. Direct access to the core. Each user-facing abstraction is a mask that can hide some ML-bug, a potential source of misunderstandings, and a steeper learning curve for users. [End]

    Prikaži ovu nit
    Poništi
  13. 14. sij

    D. Open-sourcing ML can be very different from other types of open-sourcing: - ML bugs are silent => researchers need to know exactly what's happening inside your code. - Researchers will create things you have no ideas about => they'll want to dive in your code and modify it.

    Prikaži ovu nit
    Poništi
  14. 14. sij

    7. Now if you want to build a large-scale tool like 🤗Transformers? Here are a few additional tips A. focus on one essential feature that your community really needs and no one provides B. do it well C. keep putting yourself in the shoes of people using your tool for the 1st time

    Prikaži ovu nit
    Poništi
  15. 14. sij

    5. Spend 4 days to do it well. Open-sourcing a good code base takes some time but you should consider it as important as your paper 6. Consider merging with a larger repo: are you working on language models? 🤗Transformers is probably happy to help you ➡️

    Prikaži ovu nit
    Poništi
  16. 14. sij

    3. Give clear instructions on how to run the code, at least evaluation, in such a way that, combined with pretrained models, it allows for fast test/debug 4. Use the least amount of dependencies: if you are using an internal framework to build the model => copy the relevant part

    Prikaži ovu nit
    Poništi
  17. 14. sij

    2. Put yourself in the shoes of a master student who has to start from scratch with your code: - give them a ride up to the end with pre-trained models - focus examples/code on open-access datasets (not everybody can pay for CoNLL-2003)

    Prikaži ovu nit
    Poništi
  18. 14. sij

    1. Consider sharing your code as a tool to build on more than a snapshot of your work: -other will build stuff that you can't imagine => give them easy access to the core elements -don't over-do it => no need for one-liner abstractions that won't fit other's need – clean & simple

    Prikaži ovu nit
    Poništi
  19. 14. sij

    I often meet research scientists interested in open-sourcing their code/research and asking for advice. Here is a thread for you. First: why should you open-source models along with your paper? Because science is a virtuous circle of knowledge sharing not a zero-sum competition

    Prikaži ovu nit
    Poništi
  20. proslijedio/la je Tweet

    Great intro to the modern landscape of Deep Learning & by . Including sweet mentions of the models inside transformers, write with transformers & 's NLP progress repo!

    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:

    ·