use-ml-dataset

Dataset identity and metadata primitives for RustUse machine-learning workflows.

Experimental

use-ml-dataset is experimental while use-ml remains below 0.3.0.

Example

use use_ml_dataset::{MlDatasetName, MlDatasetSplit, MlDatasetProvenance};

let name = MlDatasetName::new("iris")?;
let split: MlDatasetSplit = "validation".parse()?;

assert_eq!(name.as_str(), "iris");
assert_eq!(split.to_string(), "validation");
assert_eq!(MlDatasetProvenance::Synthetic.as_str(), "synthetic");
# Ok::<(), use_ml_dataset::MlDatasetError>(())

Scope

• Dataset names, identifiers, versions, splits, and kinds.
• Example identifiers and example-kind labels.
• Dataset license, provenance, schema-reference, and dataset-card-reference metadata.

Non-goals

• Loading datasets or parsing full dataset schemas.
• Prompt datasets, chat logs, RAG corpora, or conversation-specific concepts unless modeled generically as datasets.

License

Licensed under either Apache-2.0 or MIT.