CLIP ati Idanwo Awọn awoṣe Ede Iran

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

1. Kini ero pataki lẹhin CLIP?
   • Titẹ awọn aworan sinu awọn faili kekere
   • Gbigbe awọn aworan ati ọrọ sinu aaye ifibọ pinpin kanna
   • Ṣiṣẹda awọn awoṣe 3D lati awọn fọto
   • Itumọ ọrọ si ọrọ
2. Ọna ikẹkọ wo ni CLIP nlo?
   • Ẹkọ imudara lati awọn esi eniyan
   • Ifi aami-ọwọ gbogbo ohun ni gbogbo aworan
   • Ẹkọ itansan ti o fa awọn orisii aworan-ọrọ ti o baamu papọ
   • Didaakọ awọn iwuwo lati awoṣe ọrọ kan
3. What does zero-shot classification mean in the original CLIP transfer setup?
   • Using text-defined candidate categories without new task-specific supervised classifier fitting
   • Using a model that never received any pretraining data
   • Running without any computation or electricity
   • A guarantee that every category is classified correctly
4. Bawo ni CLIP ṣe wọn boya aworan ati akọle ibaamu?
   • Nipa lafiwe iwọn faili
   • Nipa kika awọn piksẹli ti awọ kọọkan
   • Nipa yiyewo awọn akoko ti aworan
   • Nipa ibajọra cosine laarin awọn fekito wọn
5. Ni aijọju iye data ti a ti kọ CLIP lori?
   • Awọn aworan ti a fi aami si ẹgbẹrun diẹ
   • Ọrọ nikan, ko si awọn aworan
   • Ni ayika 400 milionu awọn orisii ọrọ-aworan lati oju opo wẹẹbu
   • A nikan curated musiọmu gbigba
6. How does the cited unCLIP image-generation paper use CLIP representations?
   • The original standalone CLIP encoder directly outputs every image pixel
   • A prior supplies CLIP image embeddings to a generative decoder
   • It removes all image representations and uses only filenames
   • It uses CLIP only as a network-encryption protocol
7. Which image-encoder families were evaluated in the original CLIP paper?
   • Only recurrent text encoders
   • Only a mandatory linear pixel average
   • Only a raw audio codec
   • Modified ResNets and Vision Transformers
8. A CLIP batch contains 32 paired images and texts. How many entries are in its full cross-modal score matrix?
   • 1024
   • 32
   • 64
   • 992