Hyperparameter search

Explore independent configurations in parallel and select one with a reproducible objective.

prepare dataset→trial 1trial 2trial N→compare trials→register best

Hyperparameter search

Activity responsibilities

Activity	Responsibility
prepare-dataset	Execute the prepare dataset stage and publish its declared outputs for downstream activities.
trial-1	Execute the trial 1 stage and publish its declared outputs for downstream activities.
trial-2	Execute the trial 2 stage and publish its declared outputs for downstream activities.
trial-n	Execute the trial N stage and publish its declared outputs for downstream activities.
compare-trials	Execute the compare trials stage and publish its declared outputs for downstream activities.
register-best	Execute the register best stage and publish its declared outputs for downstream activities.

Inputs

• Dataset snapshot
• Search space
• Selection objective

Outputs

• Trial models
• Leaderboard
• Selected model

Provenance to preserve

• Parameters per trial
• Failures and retries
• Selection rationale

Runnable example

This is the complete checked-in bundle for this pattern. Download the environment, scope, topology, workflow, input, container recipe, runner, and validation contract from this page before executing it.

Read setup instructions ↓Environment YAML ↓Execution scope YAML ↓Network topology YAML ↓Workflow YAML ↓Input data ↓Container recipe ↓Run the workflow ↓Validate outputs ↓Expected output manifest ↓

Verified local execution

These captures and the output manifest were produced by the fixture's local Docker run and validator. They are published with the same bundle as the runnable files.

Workflow evidence

Execution evidence

Output evidence

Open verified output manifest ↓

Execution considerations

Cap concurrent trials by accelerator capacity, budget, or queue policy. Reuse trial identities when resuming the workflow.

AkôFlow boundary: the engine schedules, deploys, executes, transfers data, and records evidence. The ML or agent framework remains an implementation choice inside each activity.