Tilmaamaha Tubooyinka Injineernimada iyo Kediska Qaabaynta Xogta

Jawaabaha iyo sharraxaadaha waxay soo baxayaan markaad qaadanayso kediska.

1. Waa maxay ujeedada aasaasiga ah ee dhuumaha farsamada sifo?
   • Si loo kaydiyo miisaanka loo tababaray ee loo dirayo hadhow
   • Si loo jadwaleeyo shaqooyinka GPU ee kutlada
   • Si xogta ceeriin loogu beddelo sifooyinka habaysan ee moodeelku isticmaalo
   • Si aad u sirayso kaydka xogta wakhtiga nasashada
2. Why compare a dataset’s content hash instead of relying only on its filename?
   • A hash guarantees every record is accurate
   • The filename can stay the same while its stored contents change
   • A filename contains all transformations
   • Hashing removes access restrictions
3. Dhuumaha sifada ayaa inta badan lagu tilmaamaa qaab dhismehee?
   • Liis keli ah oo su'aalo SQL ah oo ay tahay in dhammaantood dib loo bilaabo mar kasta
   • Shaxda xog ururinta ee xidhiidhka ah
   • Shabakad neural ah oo leh lakabyo qarsoon
   • Garaaf acyclic toosan (DAG) ee tillaabooyinka isbeddelka
4. A model behaves differently after a data update. What can version history help investigate?
   • Which data and transformation versions were used, with other runtime factors checked separately
   • A guaranteed identical rerun without environment records
   • Proof the update was authorized
   • A universal legal-compliance certificate
5. Midkee ka mid ah aaladaha tusaale ahaan si gaar ah loogu isticmaalo dukaamada sifada ama nuqulka xogta?
   • Photoshop
   • Microsoft Eray
   • FFmpeg
   • Iid (ama DVC / LakeFS)
6. Maxay muhiim u tahay in sifo isku mid ah caqligu socdo inta lagu jiro tababarka iyo wax soo saarka labadaba?
   • Si loo dhimo qiimaha kaydinta daruuraha
   • Si moodelku u weynaado oo si toos ah u saxo
   • Si loo hubiyo in moodelku u arko sifooyin si joogto ah loo xisaabiyey oo uu ka fogaanayo tababbarka/u-adeegga khaladaadka
   • Sababtoo ah server-yada wax soo saarku ma socodsiin karaan Python
7. A team has the same data and code but a changed random state. What remains possible?
   • Different computed outputs despite matching versions
   • Versioning forces every random choice to match
   • The dataset filename must change
   • All models become deterministic
8. A reviewer asks which input record and transform produced a feature at a particular time. What record addresses that question?
   • Only the final prediction text
   • Feature lineage with source, transform version and timestamp
   • Only the dataset filename
   • Only a model parameter count