Experiments

Experiments track ML runs with parameters, metrics, and artifacts. Use this resource to create experiments, log results, compare runs, and register models.

Access it as client.experiments on a Strongly client, or the same path on AsyncStrongly with await. All methods exist on both with identical signatures.

Quick start

from strongly import Strongly

client = Strongly()

# List (auto-paginates as you iterate)  # filters: search, status, tag, pinned
for experiment in client.experiments.list():
    print(experiment.id)

# Create
created = client.experiments.create(name="lr-sweep")
print(created.id)

# Read it back
fetched = client.experiments.retrieve(created.id)

# Delete (returns None; raises on failure)
client.experiments.delete(created.id)

Methods

Core

list

list(*, search: Optional[str] = None, status: Optional[str] = None, tag: Optional[str] = None, pinned: Optional[bool] = None, limit: Optional[int] = None) -> SyncPaginator[Experiment]

List experiments with pagination, filtering, and search.

Args: search: Search by name or description. status: Filter by status. tag: Filter by tag. pinned: If True, only return pinned experiments. limit: Maximum number of items to return (default: all matching items).

create

create(*, name: str, description: Optional[str] = None, tags: Optional[Sequence[str]] = None) -> Experiment

Create a new experiment.

Args: name: Experiment name. description: Optional description. tags: Optional tags.

retrieve

retrieve(experiment_id: str) -> Experiment

Get a single experiment by ID.

update

update(experiment_id: str, *, name: Optional[str] = None, description: Optional[str] = None, tags: Optional[Sequence[str]] = None, status: Optional[str] = None) -> Experiment

Update an experiment.

delete

delete(experiment_id: str) -> None

Delete an experiment.

Other

compare

compare(ids: List[str]) -> Dict[str, Any]

Compare multiple experiments.

Args: ids: List of experiment IDs to compare (minimum 2).

log_artifact

log_artifact(experiment_id: str, *, name: str, s3_key: Optional[str] = None, content_base64: Optional[str] = None, path: Optional[str] = None, type: Optional[str] = None, content_type: Optional[str] = None, size: Optional[int] = None) -> Dict[str, Any]

Record an artifact reference for an experiment run.

Args: name: Artifact name. s3_key: S3 key if the artifact is already uploaded. content_base64: Base64-encoded file content to upload. path: Artifact path. type: Artifact type (e.g. file, model). content_type: MIME content type. size: File size in bytes.

log_metric

log_metric(experiment_id: str, *, metrics: Sequence[Mapping[str, Any]]) -> Dict[str, Any]

Append metric values to an experiment run.

Args: metrics: A sequence of {key, value, step?} metric entries.

client.experiments.log_metric(
    exp_id,
    metrics=[{"key": "accuracy", "value": 0.91, "step": 10}],
)

log_param

log_param(experiment_id: str, *, params: Mapping[str, Any]) -> Dict[str, Any]

Record hyperparameter values for an experiment run.

Args: params: Key-value parameter pairs.

client.experiments.log_param(exp_id, params={"lr": 0.01, "epochs": 20})

pin

pin(experiment_id: str) -> Dict[str, Any]

Toggle pin on an experiment.

register

register(*, name: str, description: Optional[str] = None, tags: Optional[Sequence[str]] = None) -> Dict[str, Any]

Register an experiment from an external runner.

Used by SDK consumers running training jobs outside Strongly that want their runs to land in the experiment registry.

stats

stats() -> ExperimentStats

Get experiment overview stats.

update_tags

update_tags(experiment_id: str, tags: List[str]) -> Dict[str, Any]

Update experiment tags.

Args: experiment_id: The experiment ID. tags: The new list of tags.