Seldon Core uye Inference Grafu Quiz

Mhinduro uye tsananguro zvinoonekwa paunenge uchitora bvunzo.

1. Chii chinotsanangura chinosiyanisa Seldon Core kubva kune imwechete-modhi server?
   • Inference magirafu anobatanidza akawanda mazvikamu
   • Yakavakwa-mukati modhi yekudzidziswa
   • A proprietary GPU mutyairi
   • Kunyora data otomatiki
2. Ndeipi Seldon graph chikamu chinosarudza kuti ndeipi node yemwana yekutumira chikumbiro, ichigonesa A/B bvunzo?
   • Transformer (Shanduri)
   • Router
   • Combiner
   • Model
3. Ndeipi chikamu chemhando inounganidza zvinobuda kubva kune akawanda mamodheru ekubatanidza?
   • Router
   • Transformer (Shanduri)
   • Combiner
   • Initializer
4. Ndeupi rudzi rwechimiro chegirafu chinoita Seldon inference graph fomu?
   • A cyclic mesh
   • A flat list
   • Muti wemabhinari chete
   • A directed acyclic graph (DAG)
5. MuSeldon Core v2, ndeipi sevha neprotocol inogonesa akawanda-modhi inoshumira mhiri kwegirafu?
   • MLServer uye Open Inference Protocol
   • TensorFlow Kushandira uye gRPC-chete
   • Nginx uye REST
   • Spark uye Parquet
6. In the Seldon Core 1 Router documentation, what can shift traffic toward better-performing models over time?
   • Nightly retraining only
   • A multi-armed bandit using live reward signals
   • Increasing the pod CPU limit
   • Caching every response
7. Which Seldon graph component can preprocess a request or postprocess a model response?
   • Combiner
   • Router
   • Transformer (Shanduri)
   • Explainer
8. Which description of Seldon Core 2 observability is accurate?
   • Metrics, tracing and payload logging are all automatically enabled for every REST/gRPC prediction
   • It provides only GPU metrics and no tracing
   • Payload logging always stores every model request in a local file
   • Metrics are exposed; tracing requires configuration, and payload logging is limited to configured Kafka-backed pipeline paths