Seldon Core
ML inferenceseldon-core
Kubernetes-native framework serving models, pipelines and experiments.
- Provider
- default
- Kind
- ML inference
- Category
- Compute
| property | type | values | required | kind default | this component |
|---|---|---|---|---|---|
| accelerator | enum | none | gpu | asic | no | none | inherited |
| acceleratorModel | enum | nvidia-h100 | nvidia-h200 | nvidia-b200 | nvidia-l4 | nvidia-a100 | amd-mi300x | aws-trainium | aws-inferentia | google-tpu | intel-gaudi | groq-lpu | cerebras-wse | no | — | inherited |
| embeddingSyncMinutes | integer | — | no | — | inherited |
| experiments | boolean | — | no | false | inherited |
| inferenceServer | enum | mlserver | triton | no | mlserver | inherited |
| inputs | enum | live | historical | both | no | live | ▸ live |
| instances | integer | — | no | — | inherited |
| kafkaBackbone | boolean | — | no | true | inherited |
| managed | boolean | — | no | false | ▸ false |
| pipelines | boolean | — | no | false | inherited |
| snapshotToObjectStore | boolean | — | no | false | inherited |
| training | enum | batch | online | no | batch | ▸ batch |
Every rule whose selector covers this component's kind or its category. Whether a rule fires depends on the design: its properties, its edges and its groups.
stateful-node-without-zone-redundancy
A stateful compute node pinned to exactly one availability zone disappears with that zone, and its state goes with it. Nodes inside a cell are exempt, because there the cell is deliberately single-zone and redundancy comes from the other cells.
- — all of these hold:
- — the node's `stateless` is false
- — the node sits inside exactly 1 availability-zone group
- — this does not hold:
- — the node sits inside at least one cell group
nodes:
- { id: "seldon-core", component: "seldon-core" }SageMaker endpointsagemaker-endpointAzure ML endpointazure-ml-endpointWorkers AIcloudflare-workers-aiAnyscale Endpointsanyscale-endpointsBananabanana-devBasetenbasetenBeambeam-cloudBentoMLbentomlHugging Face Inference Endpointshuggingface-inferenceInfinityinfinity-embeddingsJanjanKServekservellama.cpp serverllama-cpp-serverllm-dllm-dLM Studiolm-studioLMDeploylmdeployLocalAIlocalaiGeneric ML inferenceml-inferenceMLC LLMmlc-llmModalmodalMystic AImystic-aiNVIDIA NIM (self-hosted)nvidia-nimOctoAIoctoaiOllamaollamaRay Serveray-serveSGLangsglangTensorFlow Servingtensorflow-servingTensorRT-LLMtensorrt-llmText Embeddings Inferencetext-embeddings-inferenceText Generation Inferencetext-generation-inferenceTorchServetorchserveNVIDIA TritontritonvLLMvllmXinferencexinferenceVertex AI endpointvertex-ai-endpointOCI Data Science Model Deploymentoci-data-science-model-deploymentOCI Generative AI Dedicated AI Clusteroci-generative-ai-cluster
# yaml-language-server: $schema=../../schema/component.schema.json # Seldon Core 2: Kubernetes-native framework serving Model, Pipeline and Experiment resources. # Facts from "Seldon Core 2 Architecture": scheduler and controller control plane, Envoy data plane # routing to MLServer or Triton pods, Kafka as the backbone for Pipelines. schemaVersion: "1" type: "component" id: "seldon-core" kind: "ml-inference" provider: "default" docs: "https://docs.seldon.ai/seldon-core-2/about/architecture" defaults: training: "batch" inputs: "live" managed: false properties: inferenceServer: type: "enum" values: ["mlserver", "triton"] default: "mlserver" pipelines: type: "boolean" default: false kafkaBackbone: type: "boolean" default: true experiments: type: "boolean" default: false