quiz bu lëkkale ak tegtal · Diggante Niveau

Modèle à niveau de byte sans tokeniser

Xam xeeti làkk yu raw-byte, patch entropy BLT, ak njariñ ak njaay ci dindi vocabulaire token fixe.

Yooni njiit yu jëm ciModèle de niveau de byte gratuit tokenizer
Laaj 1 de 8

Lan la modelu niveau byte bu amul tokenizer di jàng ni ay unité de input?

Xoolaatal laaj bu nekk8 laaj

Sooy def quiz bi, tontu yi ak leeral yi dañuy feeñ.

  1. Lan la modelu niveau byte bu amul tokenizer di jàng ni ay unité de input?
    • Byte yu bees (valeur UTF-8 0 ba 255)
    • Baatu-baat yu ndaw yu bawoo ci vocabulaire bu takku
    • Baatu lëmm yuñ tàqale ay bayaal
    • embeddings yuñ njëkka xayma kese
  2. Lan mooy njariñ li gëna mag ci joxe byte yu ñor ci transformatër buñ miin?
    • Bytes mënoo màndargaal mbind bu nekkul ci Àngle
    • Toppatoo yi dañuy gëna gudd, yokk njëgu ordinatër bi
    • Barina lool valeur byte yuñ mëna samp
    • Bytes yi dina ñu ñàkk luñuy tekki ba fàww
  3. Naka lay def ba Transformateur Latent Byte (BLT) di xam fuñu wara boole ay byte ci ay patch?
    • Ni ñuy xaajalee ci bépp arafu espace
    • Jëfandikoo ab dayo patch bu 4 byte
    • Dafa sukkandikoo ci entropi biy wax luy waaja am ci niveau byte (amul wóor)
    • Randomly ci bepp paas ci kanam
  4. Why can a fixed subword vocabulary use more tokens for some languages than others?
    • Subword tokenizers only accept ASCII
    • All languages have identical token boundaries
    • They cannot represent numerals
    • Its token inventory may reflect the language distribution in the data used to build it
  5. What does removing a fixed tokenizer guarantee about byte-level input?
    • The same byte alphabet can represent text from different scripts, while learned language ability still depends on data
    • The model understands every language without training data
    • It guarantees equal performance for all scripts
    • The model no longer needs a vocabulary or learned embeddings
  6. Ban model moo njëkka wane ni ñuy doxalee ci niveau byte waaye ñu ŋàññi ko ndax dafa yeex?
    • Bert
    • ByT5
    • Word2Vec
    • GPT-2
  7. What representations does BLT’s large global Transformer process?
    • Individual words from a fixed subword vocabulary only
    • Pixel nataal
    • Dynamically sized byte patches
    • Robot joint-angle commands
  8. Which components surround the global Transformer in BLT?
    • A reward model and a value network
    • A speech vocoder and a spectrogram
    • A camera and a lidar sensor
    • A local byte encoder and a local byte decoder