VQ-VAE ak Latent yu Diskret

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

1. Lan moo wuutale espace latente bu VQ-VAE ak VAE buñ miin?
   • Dafa diskret, di jëfandikoo téere kode buñ jàng ci embeddings
   • Dafay wéy di Gaussian
   • Amul benn bayaal bu nëbbu
   • Dafay denc pixel yu ñor
2. naka la VQ-VAE di jaar ci gradient yi ci jéego kantite bu mënul wuute?
   • Dafay teggi tasaare bi ci ginaaw
   • Ak estimatër bu jub buy kopie gradient yi weesu kwantiseer bi
   • Jëfandikoo jàngat buy dooleel kese
   • Ni ñuy wërloo gradient yi ci ay lim yu mat
3. Luy 'codebook daanu'?
   • Dekodeer bi du bàyyi xel ci mbir yu nëbbu yi
   • encoder bi génne nekk lu amul àpp
   • Lu tuuti ci liñu bind ci codebook bi lañu musa jëfandikoo
   • Dayo nataal bi wàcci ba amul dara
4. Lan moo waral latente yu wuute yi am njariñ ci modeling generatif ak Transformer yi?
   • Dañuy gëna ndawal nataal yi ci disk bi kese
   • Dañu dindi soxla dekodeer
   • Dañu dindi bépp njuumte ci jamono
   • Dañuy soppi media yi ci ay token yu Transformer mëna modele ni làkk
5. Lan la VQGAN yokk ci kaw rëset bu njëkk bi ci VQ-VAE?
   • Perte adversarial ak perceptual ngir tabaxaat yu gëna ñaw
   • ab échantillonnage diffusion
   • Dindi téere kode bi
   • Encoder buy baaxoo def
6. Ban terme perte mooy tëye encoder bi 'committed' ci duggal codebook bi mu tann?
   • Diwergensi KL
   • ñàkka am engagement
   • Entropie croisé biy ñàkk
   • Triplet bi ñàkk
7. What kind of prior does the original VQ-VAE paper learn over its discrete codes?
   • A fixed prior that is never learned
   • A prior over camera locations only
   • A learned prior over the discrete code sequence
   • No prior or decoder distribution
8. Which text-to-image paper used a discrete VAE to tokenize images before autoregressive generation?
   • The original Whisper speech-recognition paper
   • A paper on Kalman-filter vehicle tracking
   • The original BERT masked-language-model paper
   • The original DALL·E paper