formatting

formatting

JIT for both labeler and detector + more drift metrics

full mention labeler inference pipeline

allow the trainer to take kwargs

fix precision oversight

fix deserialization when using labeler

add configurable stop criterion

fix: don't crash on empty predictions

add missing task-id for mention data

fix naming

fix import

entity labeler integration

integrate entity dataset

resolve

formatting and remove unused swiglu

prometheus

inference server

fix: many things

fix: joint to AND during inference

init

prometheus

inference server

fix: many things

fix: joint to AND during inference

init