llm_config: Create LLM Configuration

In LLMR: Interface for Large Language Model APIs in R

Create LLM Configuration

Description

Creates a configuration object for interacting with a specified LLM API provider.

Usage

llm_config(provider, model, api_key, troubleshooting = FALSE, ...)

Arguments

provider	A string specifying the API provider. Supported providers include: "openai" for OpenAI, "anthropic" for Anthropic, "groq" for Groq, "together" for Together AI, "deepseek" for DeepSeek, "voyage" for Voyage AI. "gemini" for Google Gemini.
model	The model name to use. This depends on the provider.
api_key	Your API key for the provider.
troubleshooting	Prints out all api calls. USE WITH EXTREME CAUTION as it prints your API key.
...	Additional model-specific parameters (e.g., 'temperature', 'max_tokens', etc.).

Value

An object of class 'llm_config' containing API and model parameters.

Examples

## Not run: 
  # OpenAI Example (chat)
  openai_config <- llm_config(
    provider = "openai",
    model = "gpt-4o-mini",
    api_key = Sys.getenv("OPENAI_KEY"),
    temperature = 0.7,
    max_tokens = 500
  )

  # OpenAI Embedding Example (overwriting api_url):
  openai_embed_config <- llm_config(
    provider = "openai",
    model = "text-embedding-3-small",
    api_key = Sys.getenv("OPENAI_KEY"),
    temperature = 0.3,
    api_url = "https://api.openai.com/v1/embeddings"
  )

  text_input <- c("Political science is a useful subject",
                  "We love sociology",
                  "German elections are different",
                  "A student was always curious.")

  embed_response <- call_llm(openai_embed_config, text_input)
  # parse_embeddings() can then be used to convert the embedding results.

  # Voyage AI Example:
  voyage_config <- llm_config(
    provider = "voyage",
    model = "voyage-large-2",
    api_key = Sys.getenv("VOYAGE_API_KEY")
  )

  embedding_response <- call_llm(voyage_config, text_input)
  embeddings <- parse_embeddings(embedding_response)
  # Additional processing:
  embeddings |> cor() |> print()

## End(Not run)

LLMR documentation built on April 4, 2025, 1:11 a.m.