Idanwo ti o sopọ mọ itọsọna · Lile Ipele

Ṣiṣan Ibamu adanwo

Explore flow matching’s learned vector fields, conditional paths, ODE sampling and relationship to diffusion-based generation.

Jẹmọ itọsọna awọn ọnaIbamu ṣiṣan
Ibeere 1 ti 8

Kini awoṣe ibaamu ṣiṣan taara kọ ẹkọ lati ṣe asọtẹlẹ?

Awotẹlẹ gbogbo ibeere8 ibeere

Awọn idahun ati awọn alaye han bi o ṣe n mu ibeere naa.

  1. Kini awoṣe ibaamu ṣiṣan taara kọ ẹkọ lati ṣe asọtẹlẹ?
    • Aaye iyara ti o gbe awọn ayẹwo lati ariwo si data
    • Aami ikasi fun aworan kọọkan
    • Awọn baiti JPEG fisinuirindigbindigbin ti aworan kan
    • Tabili wiwa ti o wa titi ti awọn iye ẹbun
  2. Bawo ni awọn ayẹwo ṣe ṣe ipilẹṣẹ lati awoṣe ibaramu sisan ti oṣiṣẹ?
    • Nipa iṣapẹẹrẹ a ọtọ Markov pq ti awọn ọrọ
    • Nipa sisọpọ aaye iyara ti o kọ ẹkọ pẹlu oluyanju ODE ti o bẹrẹ lati ariwo
    • Nipa ṣiṣiṣẹsẹhin itọsẹ lori aworan titẹ sii
    • Nipa aropin gbogbo ikẹkọ ṣeto
  3. Kini o jẹ ki ipadanu ibaamu ṣiṣan ni majemu lati ṣe ikẹkọ?
    • O foju kọ data naa patapata
    • O nilo ipinnu ni kikun aaye alapin ni itupalẹ
    • O ṣe ipo lori aaye data kan ati lo ọna interpolation ti o rọrun ti a mọ
    • O ṣiṣẹ nikan lori aworan kan ni akoko kan lailai
  4. Kini idi ti iyatọ 'sisan ti a ṣe atunṣe' jẹ ki iran yiyara ṣiṣẹ?
    • O nlo awọn nẹtiwọọki nkankikan ti o tobi julọ
    • O foju ikẹkọ patapata
    • O yi awọn aworan pada si ọrọ akọkọ
    • O ṣe atunṣe awọn ipa-ọna iṣeeṣe nitoribẹẹ awọn igbesẹ olutayo ODE diẹ ni o nilo
  5. Which named image model’s paper explicitly describes a rectified-flow transformer?
    • Stable Diffusion 3
    • Flux, as established by the cited source set
    • ResNet-50
    • YOLOv8
  6. Ni ibaramu ṣiṣan ni majemu, kini interpolation aṣoju laarin apẹẹrẹ ariwo ati ayẹwo data ni akoko t?
    • A ID Daarapọmọra ti awọn piksẹli
    • x_t = (1-t) * ariwo + t * data, idapọmọra laini taara
    • Squaring gbogbo ẹbun iye
    • Tito awọn piksẹli nipasẹ imọlẹ
  7. In conditional flow matching, how is the marginal vector field related to conditional vector fields?
    • It is the largest vector in the training batch
    • It is a fixed class label independent of time
    • It is the conditional expectation of the conditional vector field over the conditioning variable
    • It is the average pixel value of the data
  8. Beyond generative modeling, what other task does the original rectified-flow paper describe?
    • Speech diarization with speaker embeddings
    • Classifying handwritten digits from labels alone
    • Predicting calendar availability
    • Data or domain transfer between distributions