Deschide articolul original într-un tab nou Schimbarea acestei opțiuni se aplică imediat și reîncarcă pagina.
LECTURA
← Înapoi la Urmăritori

Roman Vorushin

vorushin.github.io/feed
Vizitează site-ul

Cel mai recent articol

Ultima verificare

Urmăritori 0 urmăritori
Articole: 8

  1. Research Engineers in AI labs

    (Roman Vorushin)

    I often get asked what kind of work research engineers in AI labs like Google DeepMind do.

  2. Best explanations of how LLMs work

    (Roman Vorushin)

    I maintain a set of best explanations1 of how LLMs train and work.

  3. LLM consciousness

    (Roman Vorushin)

    Yesterday Iana and I were walking through the evening hills, breathing fresh air, letting our son Robert work up his appetite before dinner, and talking.

  4. Patterns of successful AI research labs

    (Roman Vorushin)

    I stumbled upon a post of mine written a year ago, right after the loud DeepSeek R1 release.

  5. Pallas Puzzles

    (Roman Vorushin)

    tl;dr. vorushin/pallas_puzzles - JAX’s kernel language puzzles Pallas is JAX’s kernel language for writing custom operations that run on TPU.

  6. LLM pretraining on TPU v6e with a $50 budget

    (Roman Vorushin)

    Andrej Karpathy has the nanochat project with the description “The best ChatGPT that $100 can buy”.

  7. TPU/GPU collective primitives

    (Roman Vorushin)

    Training LLMs on TPUs/GPUs is often constrained not by the speed of the matrix multiplication units (MXUs), but either by data transfer from high bandwidth memory (HBM) to MXU, or by data transfer between accelerators.

  8. About me

    (Roman Vorushin)

    My name is Roman Vorushin.