download_model: Download a model manually

In localLLM: Running Local LLMs with 'llama.cpp' Backend

Download a model manually

Description

Download a model manually

Usage

download_model(
  model_url,
  output_path = NULL,
  show_progress = TRUE,
  verify_integrity = TRUE,
  max_retries = 3,
  hf_token = NULL
)

Arguments

model_url	URL of the model to download (currently only supports https://)
output_path	Local path where to save the model (optional, will use cache if not provided)
show_progress	Whether to show download progress (default: TRUE)
verify_integrity	Verify file integrity after download (default: TRUE)
max_retries	Maximum number of download retries (default: 3)
hf_token	Optional Hugging Face access token to use for this download. Defaults to the existing 'HF_TOKEN' environment variable.

Value

The path where the model was saved

Examples

## Not run: 
# Download to specific location
download_model(
  "https://example.com/model.gguf",
  file.path(tempdir(), "my_model.gguf")
)

# Download to cache (path will be returned)
cached_path <- download_model("https://example.com/model.gguf")

## End(Not run)

localLLM documentation built on Nov. 3, 2026, 1:09 a.m.