REST Endpoints

The REST endpoints are based on KServe’s v2 specification. Additional endpoints are driven by community adoption.

Health

  • GET v2/health/live: Check if the server is live

  • GET v2/health/ready: Check if the server is ready for inference requests

  • GET v2/models/{model}/ready: Check if a particular model is ready for inference requests

Metadata

  • GET v2: Get Xilinx Inference Server’s metadata

  • GET v2/hardware: Get a string describing the number and type of kernels that are available

  • GET v2/models/{model}: Get model metadata

Inference

  • POST v2/repository/models/{model}/load: Load a model. The HTML body in the response contains the endpoint to use for this model

  • POST v2/repository/models/{model}/unload: Unload a model

  • POST v2/models/{model}/infer: Make an inference request to a particular model

Observation

  • GET metrics: Get Prometheus metrics