GATE.CALIBRATE
Calibrate your own
Paste or drop ~30 labelled tickets (CSV or JSON: text,label). The on-device model scores them with the same closed-label scoring as capture; the same seeded split derives your thresholds.
Reference model: optional audit, off by default.
Adapter
Model (thresholds are per model id, never shared across models)Checking download size…
In-browser ONNX only: any causal LM with ONNX weights the transformers runtime supports, a preset, a custom id, or a dropped model folder. Architecture, size, or memory failures surface as a plain error, never a hang. Local servers and pre-scored files are the two other adapters above.
Local model folderExperimental
Drop a model folder here, or
Needs config.json, tokenizer files and single-file onnx weights. Nothing uploads anywhere.
Your labelled items · labels must be refund / review / reject
Same scoring as capture, same split as calibrate: everything above is computed from your items.