Idanwo ti o sopọ mọ itọsọna · Alabọde Ipele

LAION ati Ṣii Awọn Idanwo Datasets

Test how LAION-5B pairs, filters, and organizes web-sourced images and captions, including its CLIP embeddings and reproducibility constraints.

Jẹmọ itọsọna awọn ọnaLaion Ati Ṣii Datasets
Ibeere 1 ti 8

Kini LAION ṣe pinpin ni akọkọ ninu awọn iwe data ọrọ-aworan rẹ?

Awotẹlẹ gbogbo ibeere8 ibeere

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

  1. Kini LAION ṣe pinpin ni akọkọ ninu awọn iwe data ọrọ-aworan rẹ?
    • Awọn URL ati metadata n tọka si awọn aworan lori oju opo wẹẹbu
    • Awọn faili aworan gangan
    • Awọn akọle ọrọ nikan laisi awọn aworan
    • Iṣaju-oṣiṣẹ awoṣe òṣuwọn
  2. O fẹrẹ to awọn orisii ọrọ-aworan melo ni o wa ni LAION-5B?
    • 5 milionu
    • 5,85 bilionu
    • 500 milionu
    • 50 bilionu
  3. Awoṣe wo ni LAION lo lati ṣe àlẹmọ awọn orisii ọrọ-aworan fun didara titete?
    • GPT-3
    • Stable Diffusion
    • AGEKURU
    • ResNet-50
  4. Awoṣe ipilẹṣẹ ṣiṣi olokiki wo ni o jẹ ikẹkọ ni lilo data LAION?
    • BERT
    • AlphaFold
    • Nfọhun
    • Stable Diffusion
  5. What issue led LAION to withdraw LAION-5B in December 2023?
    • Researchers found URLs to images suspected to be child sexual abuse material
    • The dataset contained no English captions
    • The CLIP model could not read any image URLs
    • Every image in the dataset had been duplicated
  6. Awọn abajade ailera imọ-ẹrọ ti nlọ lọwọ lati LAION titoju awọn URL nikan ju awọn aworan lọ?
    • Losokepupo CLIP sisẹ
    • Asopọ rot ti o degrades reproducibility lori akoko
    • Awọn idiyele ibi ipamọ ti o ga julọ fun LAION
    • Ailagbara lati ṣe atilẹyin awọn ede pupọ
  7. Which precomputed representation can support fast similarity search across LAION-5B?
    • Full copies of every source image
    • Audio transcripts aligned to each image
    • CLIP embeddings for the image-text data
    • The complete weights of every trained text-to-image model
  8. Which three language groupings does the LAION-5B paper describe?
    • English, Spanish, and French only
    • Text-only, video, and speech datasets
    • Training, validation, and test images
    • English, multilingual, and language-undetermined data