Query Hugging Face for provider availability, pricing, latency, and
capabilities. Coverage is assembled from two sources: the Hub's
`inferenceProviderMapping`, which covers every model and task, and the router
catalogue, which adds pricing and performance metrics for chat-completion
models. Models served by no provider return an empty tibble.
Usage
hf_list_providers(model_id, token = NULL)
Arguments
- model_id
Character string. Model ID.
- token
Character string or NULL. API token for authentication.
Value
A tibble with one row per provider, including a `task` column naming
the task each provider serves.
Details
The Hub provider mapping is cached for the current R session (successful
lookups for 15 minutes). Call [hf_clear_provider_cache()] first if you need a
freshly fetched result.
Examples
if (FALSE) { # \dontrun{
# Chat model: includes pricing and latency
hf_list_providers("meta-llama/Llama-3.1-8B-Instruct")
# Task model: covered via the Hub provider mapping
hf_list_providers("BAAI/bge-small-en-v1.5")
} # }
