Kedis ku xidhan hagaha · Dhexdhexaad Heerka

Kediska Qoraal-u-Sawirka Sawirka

Dib u eeg qoraalka asalka ah ee Imagen, jiilka cascadeed, xadka firfircoon, iyo doorashada adkaanta ee tababarka xallinta sare.

Wadooyinka haga ee la xidhiidhaQoraalka Sawirka ee Sawirka
Su'aal 1 ee 8

In the original Imagen study's scaling comparison, which change produced larger image-quality and text-alignment gains?

Horudhac su'aal kasta8 su'aalo

Jawaabaha iyo sharraxaadaha waxay soo baxayaan markaad qaadanayso kediska.

  1. In the original Imagen study's scaling comparison, which change produced larger image-quality and text-alignment gains?
    • Discarding the language encoder entirely
    • Increasing the pretrained text encoder's size rather than the diffusion model's size
    • Replacing diffusion with nearest-neighbor image lookup
    • Reducing every prompt to a single word
  2. Codeeyaha qoraalka ee Imagen ku tiirsan yahay?
    • Shabakad sawir convolutional ah
    • LSTM yar
    • Xeer gacanta lagu qoray
    • Qaab luqadeed la qaboojiyey oo T5-XXL ah
  3. Sidee ayuu Imagen ku gaadhaa xallinta sare?
    • Faafinta fidsan oo leh heerar sare-u-qaadid heer sare ah
    • Tallaabo hal jiil oo weyn
    • Adoo gooyay sawir weyn oo kayd ah
    • Iyaga oo is-dhex-galiya laba sawir
  4. What problem did original Imagen's dynamic thresholding address under strong guidance?
    • A lack of text tokens in the encoder vocabulary
    • The need to compress the final PNG file
    • Saturation of predicted pixel values during sampling
    • A missing training label for every object
  5. Waa maxay DrawBench?
    • Kiniin sawir leh
    • Astaanta GPU
    • Xog ururinta sawirada bisadaha
    • Halbeegga dardargelinta adag ee qiimaynta moodooyinka qoraalka-ka-muuqaalka ah
  6. Waa maxay sababta koodhiyaha T5-XXL loogu keydiyay Imagen?
    • Way jabtay mana tababari karto
    • Fahamkeeda qoraalka qani ka ah wareejinta tababarka ka hor iyada oo aan la hagaajin dheeraad ah
    • Si loo badbaadiyo booska diskooga kaliya
    • Sababtoo ah qoraalku ma khuseeyo
  7. What resolution did original Imagen's base diffusion model generate before its super-resolution stages?
    • 256 by 256 pixels as the first stage
    • 1024 by 1024 pixels directly
    • 64 by 64 pixels
    • A single pixel that was copied across the final image
  8. Why did the Imagen researchers add Gaussian noise to low-resolution conditioning images during super-resolution training?
    • To make the upsamplers more robust to imperfect inputs from earlier stages
    • To remove text conditioning from the entire cascade
    • To guarantee that all training images had the same caption
    • To replace the upsamplers with deterministic resizing