Kediska Feejignaanta Su'aal kooxaysan

Jawaabaha iyo sharraxaadaha waxay soo baxayaan markaad qaadanayso kediska.

1. Feejignaan-kooxeed-Su'aal, maxaa la wadaagaa kooxda madaxda weydiinta?
   • Furaha iyo madaxda qiimaha
   • Saadaasha wax soo saarka kaliya
   • Ku-xidhka gelinta
   • Heerka waxbarashada
2. GQA waxaa sida ugu wanaagsan loogu tilmaamay meel dhexe oo u dhaxaysa labada daraf?
   • Convolution iyo fiiro gaar ah
   • Fiiro gaar ah oo madax badan leh iyo feejignaan su'aalo badan
   • FP32 iyo FP8 sax ah
   • Noocyo-kaliya iyo cod-dejiyaha-kaliya
3. Waa maxay qayb ka mid ah fikradda GQA ugu horrayn ka jaban tahay?
   • Calaamadaynta degdega ah
   • Xisaabinta khasaaraha tababarka
   • Xusuusta kaydka KV iyo baaxaddeeda inta lagu jiro jiilka
   • softmax ugu dambeeya ee ka sarreeya erayada
4. Haddii moodal leeyahay 32 madax waydiimo iyo 8 kooxood oo KV ah, khasnadda KV waa la dhimay qiyaastee?
   • Qiyaastii 32x
   • Qiyaastii 2x
   • Wax dhimis la'aan
   • Qiyaastii 4x
5. What quality/speed tradeoff did the GQA paper report in its experiments?
   • Quality close to MHA with inference speed closer to MQA
   • Lower quality and slower inference than both
   • Exactly the same latency and quality for every model
   • A result only on image classification
6. Sidee baa moodooyinka madax-badan ee jira inta badan loogu beddelaa GQA?
   • Adigoo tirtiraya dhammaan lakabyada dareenka
   • Talaabo gaaban oo 'kor u qaadis' ah oo laga qaaday isbaarada jirta
   • Adigoo u wareegaya calaamad kale
   • Waxba lama rogi karo
7. Which task families did the GQA paper evaluate?
   • Image segmentation and object detection
   • Speech synthesis and speaker verification
   • Language modeling and machine translation
   • Molecular generation and toxicity prediction
8. How much of the original pretraining compute did the GQA paper report using for its uptraining recipe?
   • Ilaa 50%
   • About 100%
   • No additional training compute
   • Ilaa 5%