Embeddings 0.4.0 encodedbatch

Lodestar.Embeddings 0.4.0. This page is frozen at that release. Read the current documentation for what main says now. A link to a decision or a migration page follows main, and leaves the archive.

EncodedBatch

The rectangular result: ids, attention mask, and the true lengths the padding hides.

public sealed class EncodedBatch

Properties — Count is how many sequences. SequenceLength is the width every one was padded to. InputIds holds all the ids, row-major, Count × SequenceLength of them. AttentionMask is the same shape, 1 for a real token and 0 for padding. Lengths is each sequence's true length before padding.

Example — two texts, and what the rectangle says about them.

using Lodestar.Embeddings.Tokenization;

var vocab = new Dictionary<string, int>(StringComparer.Ordinal)
{
    ["[UNK]"] = 0, ["token"] = 1, ["##ize"] = 2, ["text"] = 3,
    ["[CLS]"] = 4, ["[SEP]"] = 5, ["[PAD]"] = 6,
};
var tokenizer = new WordPieceTokenizer(
    vocab, unkToken: "[UNK]", continuationPrefix: "##", maxCharsPerWord: 100, lowercase: true);
var encoder = new BatchEncoder(tokenizer, new EncodingOptions { Template = SpecialTokenTemplate.Bert });

EncodedBatch batch = encoder.EncodeBatch(["text", "tokenize text"]);

int cells = batch.InputIds.Length;  // => 10
int first = batch.Lengths[0];       // => 3

Remarks — ten cells for two sequences of width five, three of which are padding. The attention mask is not optional: a model given padding without being told it is padding will attend to it and produce a different vector. Passing InputIds alone is the most common way to get plausible, wrong embeddings.

InputIds is row-major and flat rather than jagged, because that is the layout ONNX Runtime wants and building it here avoids a copy at the boundary.

Sequence is how to read one row without arithmetic.

Applies to — net10.0, netstandard2.0.

See also — BatchEncoder.EncodeBatch, OnnxTextEmbedder.EmbedBatch.

Members

Member	What it does
EncodedBatch.Sequence	One sequence's ids, without the padding.