This helper scans common Ollama installation directories for downloaded
GGUF weights that can be loaded directly by the `llama.cpp` backend.
It inspects both manifest metadata (when available) and the blobs directory
to return human-readable model descriptions.
Usage
list_ollama_models(min_size_mb = 50, verify = TRUE)
Arguments
- min_size_mb
Minimum size (in megabytes) for a candidate GGUF file.
Defaults to 50 MB to avoid tiny placeholder layers.
- verify
Whether to confirm the GGUF magic header before listing the
model (default `TRUE`).
Value
A data.frame with columns: `name`, `path`, `size_mb`, `size_gb`,
`size_bytes`, `sha256`, `modified`, `source`, `tag`, `model`. Returns an
empty data.frame if no models are found.