Action Recognition Quiz

Mhinduro uye tsananguro zvinoonekwa paunenge uchitora bvunzo.

1. Chii chinosiyanisa kucherechedzwa kwechiito kubva kumhando yemifananidzo yakajairika?
   • Iyo inokonzeresa nezvekuti mapixels anochinja sei nguva yese, kwete furemu imwe chete
   • Inoshandisa mifananidzo yakakura
   • Inongoshanda pavhidhiyo nhema-uye-chena
   • Izvo zvinoda hapana data yekudzidziswa
2. Muchinyakare-mbiri-rukova chiito chekucherechedza network, rwizi rwechipiri runowanzoita sei?
   • Audio waveforms
   • Optical flow inomiririra kufamba
   • Zvinyorwa zvinyorwa
   • GPS inoronga
3. Ko 3D convolution inowedzera chii kana ichienzaniswa neyakajairwa 2D convolution?
   • Nzira dzemavara
   • Higher image resolution
   • Yenguva (nguva) axis saka mafirita anotenderera akawanda mafuremu
   • Batch yakakura saizi
4. Chii chinonzi 'inflation' chinyengeri chinoshandiswa neI3D modhi?
   • Kuwedzera kupenya kwevhidhiyo
   • Kuwedzera mamwe ma layer
   • Kudzvanya vhidhiyo faira
   • Kuwedzera 2D mufananidzo-netiweki uremu mu 3D nekuadzokorora nguva dzese
5. Sei iyo Chimwe Chinhu-Chimwe dhataset ichikosha pakuzivikanwa kwechiito?
   • Inomanikidza mamodheru kuti anzwisise gwara renguva kwete kuvimba nemamiriro echiitiko
   • Iine mavhidhiyo emhuka chete
   • Haina mavara
   • Yakagadzirwa nemifananidzo imwe chete
6. Ndeipi nzira inobatsira maitiro ekuzivikanwa kuti adzidze kubva muvhidhiyo isina kunyorwa?
   • Manual frame-by-frame labeling
   • Kuzvitarisisa masiki yevhidhiyo modelling
   • Kudzima data rese rekufamba
   • Kushandisa audio chete
7. How many action-template classes did the initial Something-Something dataset paper describe?
   • 17 classes
   • 1,740 classes
   • 174 classes
   • 17,400 classes
8. What pretraining strategy does VideoMAE use?
   • Training only on single labeled frames
   • Replacing video with a text transcript
   • Predicting optical flow without reconstructing masked content
   • Masking a high proportion of video tokens and reconstructing them from visible tokens