`huggingfaceR` caches each model's Hugging Face provider mapping for the duration of the R session so that repeated inference calls do not re-query the Hub. Call this function to discard the cache after provider availability changes, or in tests.
`huggingfaceR` caches each model's Hugging Face provider mapping for the duration of the R session so that repeated inference calls do not re-query the Hub. Call this function to discard the cache after provider availability changes, or in tests.