Quiz buy soppi gis-gis

Sooy def quiz bi, tontu yi ak leeral yi dañuy feeñ.

1. How does the original plain ViT convert an image into input tokens for its encoder?
   • It must read one raw pixel per recurrent step
   • It linearly embeds flattened non-overlapping patches
   • It converts the whole image into a mandatory single scalar
   • It removes all learned input projections
2. Which mechanism supports direct interactions between allowed patch tokens in the original dense ViT encoder?
   • Fexe sa bopp
   • Only a fixed local max-pooling operation
   • Only alphabetical ordering of patch labels
   • Only a random image crop
3. Which architectural difference did the original ViT study discuss relative to convolutional networks?
   • A guarantee that it contains fewer parameters than every CNN
   • An inability to process color images
   • A guarantee that every vision task needs the same huge dataset
   • Fewer convolution-style locality and translation-equivariance biases
4. Lan mooy njariñu enkodaasu position ci Transformatëru gis-gis?
   • Ngir wàññi dayo fichier nataal bi
   • ngir jaxase patch yi ci anam wu mucc ngir yokk
   • Ngir wax model bi fi patch bu nekk nekkoon ci nataalu njëkk bi
   • Ngir soppi nataali melo yi ci grayscale
5. Benn ci njariñu bàyyi xel ci sa bopp ci patch nataal yu bari mooy njëg bi dafay yokk:
   • Ci anamu kaare ak limu patch yi
   • Lineer ak limu patch yi
   • Du dara ci dayo nataal bi
   • Ak limu chaine yu am melo yi kese
6. What conclusion is supported by the original ViT classification experiments under their stated pretraining and transfer conditions?
   • Every transformer must outperform every CNN on every dataset
   • A pure patch transformer could perform competitively with compared CNNs
   • All convolutional operations are invalid for image recognition
   • No training data is needed for image classification
7. For a 224-by-224 image divided into non-overlapping 16-by-16 patches, how many patch tokens are there before adding a class token?
   • 197
   • 224
   • 196
   • 256
8. In the original ViT classification design, which encoder state supplies the classification head?
   • Only an unembedded raw input pixel
   • A required random token not processed by the encoder
   • Only the first patch’s original filename
   • The final state of the learned class token