Crate privacy_filter_rs 

privacy-filter-rs — OpenAI Privacy Filter inference in Rust

Pure-Rust inference for the OpenAI Privacy Filter token classification model, built on Burn 0.20.

Quick start

use privacy_filter_rs::PrivacyFilterInference;
use std::path::Path;

// Choose backend
type B = burn::backend::NdArray;
let device = burn::backend::ndarray::NdArrayDevice::Cpu;

let engine = PrivacyFilterInference::<B>::load(
    Path::new("/path/to/privacy-filter"),
    device,
)?;

let spans = engine.predict("My name is Alice Smith")?;
for span in &spans {
    println!("{}: {} (score: {:.4})", span.entity_group, span.word, span.score);
}

Re-exports

pub use config::ModelConfig;

pub use config::ViterbiConfig;

pub use inference::PrivacyFilterInference;

pub use viterbi::PrivacySpan;

Modules

backend

config

inference

model

tensor_utils

viterbi

weights

Functions

init_threads

Configure the global Rayon thread pool.