Module llm

Client implementation for interacting with Ollama LLMs.

This module provides functionality to interact with locally running Ollama LLMs, particularly focused on the Llama model family. It supports various API endpoints including chat, generation, and embeddings, with primary support for chat-based interactions.

The client handles URL management, request building, and response parsing while providing sensible defaults and helpful warnings when using fallback configurations.

Examples

use learner::llm::{LlamaRequest, Model, OllamaEndpoint};

let request = LlamaRequest::new()
  .with_host("http://localhost:11434")
  .with_endpoint(OllamaEndpoint::Chat)
  .with_model(Model::Llama3p2c3b)
  .with_message("What is quantum computing?");

let response = request.send().await?;
println!("Response: {}", response.message.content);

Structs

• LlamaRequest
  Request builder for Ollama LLM interactions.
• LlamaResponse
  Response structure from Ollama LLM requests.
• Message
  Message structure for LLM interactions.
• Options
  Configuration options for LLM inference.

Enums

• Model
  Available models for use with the Ollama service.
• OllamaEndpoint
  Available API endpoints for the Ollama service.