Pix2Pix Aworan-si-Aworan Idanwo Itumọ

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

1. Iru data ikẹkọ wo ni Pix2Pix nilo?
   • Awọn apẹẹrẹ aworan igbewọle-jade ti a so pọ
   • Ko si data rara
   • Awọn akọle ọrọ nikan
   • Awọn igbasilẹ ohun
2. Kini faaji monomono ti Pix2Pix lo?
   • Nẹtiwọọki ti a ti sopọ ni kikun
   • A U-Net pẹlu foo awọn isopọ
   • Nẹtiwọọki loorekoore
   • Igi ipinnu
3. Kini iyasọtọ PatchGAN ni Pix2Pix ṣe iṣiro?
   • Gbogbo aworan bi ọkan agbaye Dimegilio
   • Iwọn faili aworan naa
   • Otitọ ti awọn abulẹ aworan agbegbe kekere
   • Akoko ikẹkọ
4. Kini idi ti Pix2Pix ṣe ṣafikun pipadanu L1 lẹgbẹẹ ipadanu ọta naa?
   • Lati fa fifalẹ ikẹkọ lori idi
   • Lati yọ iyasoto kuro
   • Lati se ina ID ariwo
   • To anchor pixelwise and low-frequency fidelity while the adversarial loss encourages sharper texture
5. Kini idi ti awọn asopọ foo U-Net ṣe iranlọwọ paapaa fun awọn iṣẹ ṣiṣe itumọ?
   • Wọn kọja eto pinpin bi awọn egbegbe taara kọja nẹtiwọọki naa
   • Wọn pa awọn ipele ti ko wulo
   • Wọn yi awọn aworan pada si ọrọ
   • Wọn dinku iwọn dataset
6. Awoṣe wo ni o gbooro si Pix2Pix lati kọ ẹkọ lati awọn akojọpọ aworan ti a ko so pọ?
   • ResNet
   • CycleGAN
   • BERT
   • Ọrọ2Vec
7. Why can a PatchGAN discriminator encourage sharper local image texture?
   • It predicts the input labels directly
   • It replaces the generator’s U-Net
   • It judges realism in small image patches rather than only scoring the whole image
   • It computes only global color histograms
8. In Pix2Pix, which part of the objective most directly anchors an output to its paired target image?
   • The learning-rate warmup
   • The GAN’s random seed
   • The PatchGAN receptive-field count
   • The L1 pixel-difference loss