Classify Text (HuggingFace)

@huggingface1 VersionClassify

Runs a fine-tuned text-classification model over a String and emits ranked [Classification], a single Classification, or raw [Double] scores in the model's label order, by output slot, when a trained model exists (sentiment, topic, intent).

How it fits

Typical backends

Sentiment routing on transcribed speech, published as JSON.

Toxicity gate where the top label drives a branch before logging.

Caveats

Parameter interactionhf_model_id, device, and function_to_apply are hot-swappable at runtime; top_k and threshold are re-read every message.
I/O contractThe model's own label set drives the output; on the [Double] arm scores align to the model's label-id order, so swapping hf_model_id for a different label set changes the vector meaning.
Parameter interactiontop_k and threshold apply only to the ranked [Classification] arm; the single-best arm returns the top label and the [Double] arm returns every score.
AccuracyInputs longer than the model's context window are truncated; long documents should be chunked upstream when every section matters.
FallbackEmpty or whitespace-only String input returns the empty form of the resolved arm without running the model; downstream must handle the zero-element case.

Versions

11798943latestdefault
Automated release
6/25/2026