Baatu Glossaire IA

Luy Piyeesu frase?

Maanaam

Tokenizer bu xamul làkk, di jàng ay subword ci bind bu ñor te duko njëkka xaaj ci espace bu weex.

Term yu ci méngoo

KLIP
Architecture model multimodal buy jàng misaal yuñ bokk ci diggante bind ak nataal.
Wéyal jàng
Xeetu tàggat yaram buy may model bi mu wéy di jàng ci done yu bees te du fàtte xam-xam bimu amoon bu njëkk.
Jàng ci biir muy tënk
Mbaaxu model bi ci topp motif yi ci misaal yiñ joxe ci laaj bi.
Xeetu mébet
Xalaatal jubluwaayu jëfandikukat bi ci ab laajtu bind ngir yóbbu ko ci anam wu jaar yoon.
Modelu làkk bu mag (LLM)
Benn xeetu làkk buñ tàggat ci corpus mbind yu bari ngir sos ak jàngat mbind.
Jeton
Benn pàccu mbind buñ toppatoo ci misaali làkk, lu ci melni benn baat wala màndarga.

Nekk leen ci sunuy tegtal yu amul fayda

SentencePiece Tokenization

Xoolal itam

BPE (Byte Pair Encoding)ANN (Approximate Nearest Neighbor)Sliding Window AttentionHNSWALiBiRerankerRoPE (Rotary Positional Embeddings)Hybrid Search
Xoolal Glossaire IA bi yéppSaytu bépp tegtal IA bu amul fayda