Ikibazo gihuza ibibazo · Biragoye Urwego

Byihuse-Kuri-Byihuta Kwambukiranya-Kwandika Ikibazo

Shakisha uburyo Byihuse-Kuri-Byihuta byahinduye amashusho binyuze muguhuza ibitekerezo, guhuza aho, no guhinduranya kumafoto nyayo.

Inzira zijyanye no kuyoboraByihuse Kwihutisha Kwitondera Guhindura
Ikibazo 1 Bya 8

Ni ayahe makuru y'imbere yihutira gukoresha kugirango abungabunge ibintu?

Reba buri kibazo8 ibibazo

Ibisubizo nibisobanuro bigaragara mugihe ufata ikibazo.

  1. Ni ayahe makuru y'imbere yihutira gukoresha kugirango abungabunge ibintu?
    • Ikarita-ihuza amakarita ahuza amagambo n'uturere twamashusho
    • Ibisohoka byanyuma pigiseli
    • Icyitegererezo cyiza
    • Ishusho yo kwikuramo metadata
  2. Ese Kwihuta-Kuri-Gusaba bisaba guhuza neza icyitegererezo?
    • Nibyo, imyitozo yuzuye kuri buri guhindura
    • Oya, ni tekinike idafite amahugurwa
    • Nibyo, ariko kodegisi yinyandiko gusa
    • Gusa kugirango wongere amagambo mashya
  3. Guhinduranya 'igare' kuri 'moto' mugihe ukomeje kwerekana, uburyo:
    • Menyesha akarere k'amagare n'intoki
    • Hugura LoRA nshya kuri moto
    • Ikarita yitondere hagati yikimenyetso kijyanye nibisobanuro byombi
    • Bitanga ishusho yigenga rwose
  4. Kongera uburemere bwikimenyetso bikora iki?
    • Irahindura imiterere yishusho
    • Isiba ikimenyetso rwose
    • Ihinduranya amagambo mubisubizo
    • Irashimangira cyangwa igabanya iryo jambo imbaraga zo kureba
  5. Iyo wongeyeho ijambo kubibazo, Byihuse-Kuri-Byihuse:
    • Komeza kwitondera ibimenyetso bidahindutse kandi ureke gusa ikimenyetso gishya kigire ibitekerezo bishya
    • Ongera uhindure ishusho yose
    • Kwirengagiza ijambo ryongeweho
    • Hindura muburyo butandukanye
  6. Which separate model was trained to follow written image-editing instructions from an input image?
    • InstructPix2Pix, a conditional diffusion editor trained on generated instruction-edit pairs
    • Prompt-to-Prompt, which edits by controlling cross-attention maps during denoising
    • Null-text inversion, which optimizes unconditional embeddings for image inversion
    • LocalBlend, which creates a selected-token mask from attention maps
  7. What does the Prompt-to-Prompt LocalBlend feature use to localize a change?
    • A hand-drawn mask required for every Prompt-to-Prompt edit
    • The EXIF metadata of the original image
    • Attention maps for selected prompt words to construct a spatial mask
    • A classifier that identifies image copyright
  8. What extra step can enable Prompt-to-Prompt editing of a real photograph?
    • Train a new diffusion model from scratch
    • Convert the photograph into an audio spectrogram
    • Replace every cross-attention layer with a classifier
    • Invert the image into the diffusion model’s latent trajectory first