MbppDataLoader<T>

Models & Types · AiDotNet.Data.Text.Benchmarks

Loads the MBPP (Mostly Basic Python Problems) benchmark — 1,000 entry-level Python coding problems with natural-language descriptions and unit tests (Austin et al. 2021).

Inheritance: DataLoaderBase<T> → InputOutputDataLoaderBase<T, Tensor<T>, Tensor<T>> → MbppDataLoader<T>

How It Works

Expects {DataPath}/mbpp.jsonl. Auto-download fetches the canonical Google research GitHub release. Each record has text (problem), code (canonical solution), test_list, and task_id. Standard splits use task_id ranges: 1–10 prompts (skipped here), 11–510 test, 511–600 val, 601–974 train.

Constructors

Constructor Summary
MbppDataLoader(MbppDataLoaderOptions options) Creates a new MBPP data loader.

Properties

Property Summary
int FeatureCount { get; }
int OutputDimension { get; }
int TotalCount { get; }
string Description { get; }
string Name { get; }

Methods

Method Summary
ValueTuple<IInputOutputDataLoader<T, Tensor<T>, Tensor<T>>, IInputOutputDataLoader<T, Tensor<T>, Tensor<T>>, IInputOutputDataLoader<T, Tensor<T>, Tensor<T>>> Split(double trainRatio, double validationRatio, int? seed)