Quiz buy saytu lu nëbbu ci bopp yu bari

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

1. Lan mooy jafe-jafe bi gëna mag bi ñu nara wàññi?
   • Njëg li ci mémoire bi ci cache valeur-clé bi ci diiru inference
   • Done yu néew ci wàllu tàggat yaram
   • Limu paramet model yi
   • Dayo vokabileer tokenizer
2. naka la MLA di wàññi cache KV bi?
   • Dafay efaase jeton yu yàgg yi ci cache bi
   • Dafay denc ab vecteur latent bu ndaw bu am rang bu woyof ci palaasu caabi ak valeur yu mat
   • Dafay jëfandikoo poids 1-bit ci lépp
   • Dafay sànni bàyyi xel ci yenn jeton
3. Ban model moo njëkka dugal xel mu nëbbu ci bopp yu bari?
   • GPT-2
   • Bert
   • Seetlu bu xóot-V2
   • Llama 1
4. Lan moo tax MLA soxla yoon wu wuute 'decouplé' ngir position rotary?
   • Ndax position yi amul benn njariñ ci bàyyi xel
   • Ndax RoPE dafay yokk vocabulaire bi ñaari yoon
   • Ndax dafay gaawlu optimisatër bi
   • Ndax wëréel gi mënul nekk ci projection biy absorbé ni yeneen caabi yi mëna defee
5. Lu tollu ci ñaata KV-cache memory la DeepSeek-V2 xamle ci MLA ak bàyyi xel ci bopp yu bari?
   • Lu ëpp 90%
   • Lu tollu ci 5%
   • 50% ci dëgg
   • Amul wàññi, gaawaay kese
6. Lan mooy njariñ li ci topp ci cache bu gëna ndaw bu MLA ci token bu nekk?
   • dindi soxla bépp GPU
   • Dayo inference yu gëna mag ak contexte yu gëna gudd ci benn aparey
   • Tekki làkk yépp ci boppam
   • Dindi jëfandikoo softmax
7. What is the purpose of absorbing MLA up-projection weights into other projections?
   • Remove positional information from every token
   • Increase the size of the cached latent
   • Avoid materializing full per-head keys and values during inference
   • Replace attention with a recurrent network
8. What positional information does MLA handle on a separate decoupled key path?
   • Class labels from the training data
   • Image segmentation maps
   • The output vocabulary probabilities
   • Rotary position embeddings that cannot be absorbed like the other key projections