Re-exports§
Modules§
Structs§
- Candidate
Predictor - An iterator that predicts a sequence of tokens until the end of the sequence is reached.
- Candidates
- A container for candidate tokens.
- Engine
- An
Engineencompasses everything needed to run inferences. It contains the model and the context. It is the main entry point for running inferences. - Invalid
Probability - Error for invalid probability values.
- Model
- An ergonomic wrapper for a
llama.cppmodel. - NGram
- An immutable N-gram of tokens.
- NGram
Data - Metadata about an Ngram.
- NGram
Stats - A map of
NGrammetadata. - Piece
Predictor - A predictor that predicts pieces of text.
- Predict
Options - Options for prediction.
- Predicted
- Contains a token and the associated piece. This is a convenience struct to avoid ackward iterator usage when both the token and piece are needed.
- Predictor
- Probability
- A
Probabilityis a wrapper around a floating point number that represents a probability. It is guaranteed to be between 0.0 and 1.0. - Repetition
Options - Options for
apply_sample_repetition_penalties. - Sample
Options - Options determining how raw logits are turned into a token. This is used by
Candidates::sample_tokenand associated functions. - Token
Data Array - Token
Predictor - Vocab
Enums§
- NewError
- Possible errors when creating a new
Engine. - Repetition
Error - Sampling
Mode - Sorted
- Sort state of the candidates.
- Vocab
Kind
Constants§
Functions§
- llama_
quantize - Quantize a Llama model.