infer

Classify text input into semantic types — single values, files, or streaming batches.

Classify text into one of FineType's 250 semantic types. Use infer to explore individual values, process files of inputs, or stream column batches via JSONL.

Usage

finetype infer [OPTIONS]

Options

Flag	Type	Default	Description
`-i`, `--input`	string	—	Single text input to classify
`-f`, `--file`	path	—	File containing inputs (one per line)
`-m`, `--model`	path	`models/default`	Model directory
`-o`, `--output`	string	`plain`	Output format: `plain`, `json`, `csv`, `markdown`, `arrow`
`--confidence`	flag	—	Include confidence score in output
`-v`, `--value`	flag	—	Include the input value in output
`--model-type`	string	`char-cnn`	Model type: `transformer`, `char-cnn`, `tiered`
`--mode`	string	`row`	Inference mode: `row` or `column`
`--sample-size`	integer	`100`	Sample size for column mode
`--bench`	flag	—	Print throughput statistics to stderr
`--header`	string	—	Column name for header hint (column mode)
`--batch`	flag	—	Read JSONL from stdin (requires `--mode column`)
`--sharp-only`	flag	—	Disable Sense classifier (Sharpen-only pipeline)

Examples

Classify a single value

$ finetype infer -i "[email protected]" --confidence
identity.person.email	1.0000

Classify several values

$ finetype infer -i "192.168.1.10" --confidence
technology.internet.ip_v4	1.0000

$ finetype infer -i "2024-01-15T09:30:00Z" --confidence
datetime.timestamp.iso_8601	1.0000

$ finetype infer -i "not-an-email" --confidence
identity.person.username	0.9963

Classify from a file

Given a file samples.txt with one value per line:

$ finetype infer -f samples.txt -o json --confidence

Each line is classified independently in row mode (the default). Row mode answers: "what is this single value?"

Column mode

Column mode classifies a set of values as a group, answering: "what type does this column contain?" This is what profile uses internally.

$ finetype infer -f column_values.txt --mode column --header "email" --confidence
identity.person.email	1.0000

Providing --header gives the model a hint from the column name, improving accuracy for ambiguous values.

Row mode vs column mode

	Row mode	Column mode
Input	Single value	Set of values
Question answered	"What is this value?"	"What type is this column?"
Header hints	Not applicable	Optional via `--header`
Use case	Exploration, debugging	Profiling, batch classification