Kedis ku xidhan hagaha · Dhexdhexaad Heerka

Kediska Models Multimodal Reka AI

Baro sida nidaamyada hab-dhaqameedka badan ee Reka u habeeyaan sawirada iyo fiidyaha, iyo sida moodallada Edge loogu talagalay geynta gudaha iyo geesaha.

Wadooyinka haga ee la xidhiidhaQaababka badan ee Reka AI
Su'aal 1 ee 8

Which input types did Reka’s 2024 Core, Flash, and Edge technical report describe?

Horudhac su'aal kasta8 su'aalo

Jawaabaha iyo sharraxaadaha waxay soo baxayaan markaad qaadanayso kediska.

  1. Which input types did Reka’s 2024 Core, Flash, and Edge technical report describe?
    • Text, images, video, and audio
    • Only text and spreadsheet cells
    • Only still images
    • Audio and source code but no visual input
  2. What does Reka say its Vision system can do with video and image content?
    • Compile it into executable code
    • Search, interpret, and reason across the content
    • Guarantee that every event is detected correctly
    • Convert it only into a spreadsheet
  3. Which Reka model is described as an open-weight 7B vision-language model for deployment at the edge?
    • Reka Core
    • Reka Vision
    • Reka Edge
    • Reka Flash
  4. What did Reka’s 2024 technical report say about how its models could be deployed?
    • They could only run in a public cloud
    • They were limited to a browser extension
    • They could only run on an audio recorder
    • It described deployment flexibility, including on-premises or on-device options
  5. What does Reka’s current API model-list endpoint return?
    • The model IDs available to the authenticated user
    • A guaranteed list of every model ever released
    • Training data for each model
    • Only benchmark rankings
  6. How should readers interpret Reka’s published model performance comparisons?
    • As guarantees for every device and workload
    • As results tied to the evaluated models, tasks, and test conditions
    • As proof that every model is equally capable
    • As a substitute for checking current model documentation
  7. How did Reka’s 2025 visual-language model description handle video frames?
    • It converted the whole video into one audio waveform
    • It processed only the file name and ignored the frames
    • It sampled frames and combined their visual features with text and audio features
    • It replaced all video input with handwritten labels
  8. What factors does Reka say affect actual latency and memory use for an Edge deployment?
    • Kaliya magaca suuqgeynta moodeelka
    • The color space of the input photo
    • The number of API users in an unrelated account
    • The model, runtime, and device