from_exp

classmethod Cluster.from_exp(exp, cluster_config=None, data_dir=None)

Initialise from experiment for clustering of entire data that went into creating training data.

Cluster parameters are read from Experiment.config['cluster'].

Parameters

• exp (Experiment) – Trained Experiment.

• cluster_config (Mapping[str, Any] | None) – Additional cluster parameters (like subsampling etc). Overwrites default cluster parameters.

• data_dir (str) – Directory containing cluster_mpp (relative to {exp.dir}/{exp.name}).

Return type

Cluster