RWKV Laaj ak tontu ci ligneer

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

1. Lan mooy moomeel bu mag bi ci RWKV?
   • Dafay tàggat ni Transformer waaye dafay daw ni RNN
   • Mën nañu ko tàggat ci nataal kese
   • Dafay laaj ab cache valeur-caabi quadratique
   • Mënu ñu ko méngale dara
2. Lu arafu RWKV yi di tekki?
   • Liggéey, bind, denc, xool
   • Reception, Poids, Clé, Valeur
   • Recurrent, Yaatu, Kernel, Vecteur
   • Random, Pondéré, Xam-xam, Coppite
3. naka lay def ba RWKV mëna wéy di am mémoire ci diiru generation bi?
   • Soo efaasee poid model yi
   • Soo denc bépp token ci disk bi
   • Ni ñuy yóbbu ab etaa bu nëbbu bu ndaw bu am dayo bu takku ci kanam ci barabu ab cache buy màgg
   • Ni ñuy tënk génne yi ci fukki jeton
4. Ban cër la poids de décay temps (W) di def?
   • Dafay tëral ni ñuy jàngee
   • Dafay lim limu jeton yi
   • Dafay tànn làkku génne
   • Dafay saytu ni xibaar bu yàgg bi di gaawe ñàkka am doole
5. Buñu ko méngale ak bàyyi xel ci softmax, lan la RWKV di moytu ci bàyyi xel ci ligneer?
   • Tegtale token quadratic
   • Bépp jëfandikoo caabi ak valeur
   • Soxla embedding
   • Jëfandikoo reso neuronal
6. Lan moo waral RWKV gëna méngoo ak defar streaming?
   • Ndax fàww mu gis duggal gi yépp balaa génne gu nekk
   • Ndax token bu bees bu nekk mingi aju ci stade fixe-size bi njëkk
   • Ndax mënul defar mbind
   • Ndax dafa soxla GPU cluster ngir token bu nekk
7. In the guide’s RWKV description, what does the receptance gate control?
   • How many image patches are generated
   • The size of the model vocabulary
   • How much of the accumulated state is read out
   • The amount of GPU memory available
8. What training property does RWKV retain from Transformers?
   • Full quadratic key-value caching at inference
   • A separate generator and discriminator
   • A fixed top-k token router in every layer
   • Parallelizable training across sequence positions