Skip to main content

​
Recipes and Inference Lab

Recipes turn one real deployment revision into a content-addressed, reproducible record. InferCrane will only capture a recipe when the active revision has an immutable ModelArtifact, an explicit runtime version, and a matching persisted AIPerf benchmark.
The recipe payload records the immutable model commit, runtime/version/arguments, provider adapter, region, GPU, compute mode, workload contract, benchmark workload, and benchmark ID. Provenance records the deployment/revision, artifact, benchmark tool/version, InferCrane version, and evidence class. The SHA-256 digest covers both canonical objects. Retrying the same name/version is idempotent only for identical content; different content is rejected.

​
Compare evidence

Inference Lab v1 compares existing benchmark history. It does not provision hardware or run another load generator:
Rows are ordered by measured p95 TTFT and include runtime, provider, GPU, error rate, throughput, SLO result when measurable, and cost metadata exactly as persisted. Current Lab results emit only MEASURED rows. The schema reserves modeled and heuristic labels for future evidence sources, but InferCrane never substitutes them silently. With no comparable benchmark, the result is an empty persisted evaluation. Use --workload-digest to require an exact canonical benchmark workload. Prompt and output content are not part of recipes, benchmark history, or Lab evaluations.

​
Qualification boundary

A captured recipe proves provenance, not universal portability. Provider/runtime support remains scoped to the integration matrix, and local benchmark fixtures are not real GPU evidence. Cost remains unavailable unless a trustworthy source and observation timestamp were persisted with the benchmark.