Package index • localLLM

Package Overview

localLLM-package localLLM: R Interface to llama.cpp with Runtime Library Loading

Quick Start

Simple functions to get started quickly

quick_llama(): Quick LLaMA Inference

quick_llama_reset(): Reset quick_llama state

Core Functions

Main text generation and annotation functions

generate(): Generate Text Using Language Model Context

generate_parallel(): Generate Text in Parallel for Multiple Prompts

explore(): Compare multiple LLMs over a shared set of prompts

Model Management

Functions for loading and managing models

model_load(): Load Language Model with Automatic Download Support

context_create(): Create Inference Context for Text Generation

download_model(): Download a model manually

list_cached_models(): List cached models on disk

list_ollama_models(): List GGUF models managed by Ollama

Backend Setup

Installation and backend management

install_localLLM(): Install localLLM Backend Library

backend_init(): Initialize localLLM backend

backend_free(): Free localLLM backend

lib_is_installed(): Check if Backend Library is Installed

get_lib_path(): Get Backend Library Path

Tokenization

Token manipulation functions

tokenize(): Convert Text to Token IDs

detokenize(): Convert Token IDs Back to Text

tokenize_test(): Test tokenize function (debugging)

Chat Templates

Functions for formatting chat conversations

apply_chat_template(): Apply Chat Template to Format Conversations

smart_chat_template(): Smart Chat Template Application

apply_gemma_chat_template(): Apply Gemma-Compatible Chat Template

Reliability & Validation

Annotation reliability and validation tools

intercoder_reliability(): Intercoder reliability for LLM annotations

compute_confusion_matrices(): Compute confusion matrices from multi-model annotations

validate(): Validate model predictions against gold labels and peer agreement

Utilities

Helper functions and utilities

annotation_sink_csv(): Create a CSV sink for streaming annotation chunks

hardware_profile(): Inspect detected hardware resources

get_model_cache_dir(): Get the model cache directory

set_hf_token(): Configure Hugging Face access token

document_start(): Start automatic run documentation

document_end(): Finish automatic run documentation

Data

Example datasets

ag_news_sample: AG News classification sample