public static interface GkeInferenceQuickstartGrpc.AsyncServiceGKE Inference Quickstart (GIQ) service provides profiles with performance metrics for popular models and model servers across multiple accelerators. These profiles help generate optimized best practices for running inference on GKE.
Methods
fetchBenchmarkingData(FetchBenchmarkingDataRequest request, StreamObserver<FetchBenchmarkingDataResponse> responseObserver)
public default void fetchBenchmarkingData(FetchBenchmarkingDataRequest request, StreamObserver<FetchBenchmarkingDataResponse> responseObserver)Fetches all of the benchmarking data available for a profile. Benchmarking data returns all of the performance metrics available for a given model server setup on a given instance type.
| Parameters | |
|---|---|
| Name | Description |
request |
FetchBenchmarkingDataRequest |
responseObserver |
io.grpc.stub.StreamObserver<FetchBenchmarkingDataResponse> |
fetchModelServerVersions(FetchModelServerVersionsRequest request, StreamObserver<FetchModelServerVersionsResponse> responseObserver)
public default void fetchModelServerVersions(FetchModelServerVersionsRequest request, StreamObserver<FetchModelServerVersionsResponse> responseObserver) Fetches available model server versions. Open-source servers use their own
versioning schemas (e.g., vllm uses semver like v1.0.0).
Some model servers have different versioning schemas depending on the
accelerator. For example, vllm uses semver on GPUs, but returns nightly
build tags on TPUs. All available versions will be returned when different
schemas are present.
| Parameters | |
|---|---|
| Name | Description |
request |
FetchModelServerVersionsRequest |
responseObserver |
io.grpc.stub.StreamObserver<FetchModelServerVersionsResponse> |
fetchModelServers(FetchModelServersRequest request, StreamObserver<FetchModelServersResponse> responseObserver)
public default void fetchModelServers(FetchModelServersRequest request, StreamObserver<FetchModelServersResponse> responseObserver) Fetches available model servers. Open-source model servers use simplified,
lowercase names (e.g., vllm).
| Parameters | |
|---|---|
| Name | Description |
request |
FetchModelServersRequest |
responseObserver |
io.grpc.stub.StreamObserver<FetchModelServersResponse> |
fetchModels(FetchModelsRequest request, StreamObserver<FetchModelsResponse> responseObserver)
public default void fetchModels(FetchModelsRequest request, StreamObserver<FetchModelsResponse> responseObserver) Fetches available models. Open-source models follow the Huggingface Hub
owner/model_name format.
| Parameters | |
|---|---|
| Name | Description |
request |
FetchModelsRequest |
responseObserver |
io.grpc.stub.StreamObserver<FetchModelsResponse> |
fetchProfiles(FetchProfilesRequest request, StreamObserver<FetchProfilesResponse> responseObserver)
public default void fetchProfiles(FetchProfilesRequest request, StreamObserver<FetchProfilesResponse> responseObserver)Fetches available profiles. A profile contains performance metrics and cost information for a specific model server setup. Profiles can be filtered by parameters. If no filters are provided, all profiles are returned. Profiles display a single value per performance metric based on the provided performance requirements. If no requirements are given, the metrics represent the inflection point. See Run best practice inference with GKE Inference Quickstart recipes for details.
| Parameters | |
|---|---|
| Name | Description |
request |
FetchProfilesRequest |
responseObserver |
io.grpc.stub.StreamObserver<FetchProfilesResponse> |
generateOptimizedManifest(GenerateOptimizedManifestRequest request, StreamObserver<GenerateOptimizedManifestResponse> responseObserver)
public default void generateOptimizedManifest(GenerateOptimizedManifestRequest request, StreamObserver<GenerateOptimizedManifestResponse> responseObserver)Generates an optimized deployment manifest for a given model and model server, based on the specified accelerator, performance targets, and configurations. See Run best practice inference with GKE Inference Quickstart recipes for deployment details.
| Parameters | |
|---|---|
| Name | Description |
request |
GenerateOptimizedManifestRequest |
responseObserver |
io.grpc.stub.StreamObserver<GenerateOptimizedManifestResponse> |