Jev vs GLiNER2.5 benchmark

A framework comparing hosted Jev with a local GLiNER2.5 on zero-shot single-label classification (AG News, Banking77, DAIR Emotion) for accuracy, calibration, selective-automation coverage and latency, with bootstrap intervals.

Notes

Pilot on 300 examples: Jev 0.910 / 0.870 / 0.480 versus GLiNER 0.700 / 0.610 / 0.440; Jev 236–296 ms hosted versus ~44 ms local for GLiNER.

source-backed — Public repository, docs or live artifact. About this label

Destinations

Filed under

• #benchmark
• #classification
• #github

Also on Jevland by AbdelStark

• awesome-typesafe directory · source-backed

In collections

• Benchmarks & independent tests →