README.md

metadata

language:
  - en
  - de
  - fr
  - es
  - pt
  - it
  - pl
  - ru
  - zh
  - ja
  - ko
  - ar
  - hi
  - id
  - vi
  - th
tags:
  - liquid
  - lfm2
  - lfm2.5
  - bidirectional
  - masked-lm
  - encoder
  - pii
  - ner
  - privacy
  - multilingual
  - token-classification
library_name: transformers
license: other
license_name: lfm1.0
license_link: LICENSE
pipeline_tag: token-classification
base_model:
  - LiquidAI/LFM2.5-Encoder-350M

Try LFM • Docs • LEAP • Discord

LFM2.5-Encoder-350-PII-Detector

A full fine-tune of LFM2.5-Encoder-350M with a token-classification head, covering 40 PII types across 16 languages (en, de, fr, es, pt, it, pl, ru, zh, ja, ko, ar, hi, id, vi, th). Ships with an inference-timemhybrid regex decode (pii_hybrid_decode.py) that adds validator-gated formats (email/IBAN/credit-card/IP/JWT/…) and cue-gated IDs on top of the model.

Trained on a persona-driven, gemma-generated synthetic corpus (coherent locale-personas × scenarios × cue/inline/structured embedding × ID-contrastive disambiguation), LLM-judge-filtered and contamination-cleaned against all evaluation sets.

Find more details about our encoders in our blog post.

💻 Demos: Try this fine-tuned model running in a CPU-only Hugging Face space: PII detection** — spot and remove 40 kinds of personal information across 16 languages.

Entity types (40 PII types across 11 domains)

Domain	Types
Identity	identity.person_name, identity.ssn, identity.national_id, identity.passport, identity.drivers_license, identity.date_of_birth, identity.tax_id
Contact	contact.email, contact.phone, contact.address, contact.postal_code, contact.ip_address
Financial	financial.credit_card, financial.iban, financial.bank_account, financial.swift_bic, financial.crypto_wallet, financial.amount
Credentials	credential.api_key, credential.password, credential.private_key, credential.jwt, credential.connection_string, developer.login_credentials
Online	online.username, online.url
Device	device.mac_address, device.imei, developer.device_id
Location	location.gps_coordinates
Healthcare	healthcare.medical_record, healthcare.condition, healthcare.medication, healthcare.health_plan_id
Organization	org.company_name
Special-category	special.religion, special.political, special.orientation, special.health_status
Legal	legal.case_number

Benchmarks (18-locale-filtered, partial-F1, hybrid decode)

Benchmark	this model	detection-tier	prev (v8)	GLiNER	LFM-demo-q4
SPY	0.428	0.509	0.351	0.280	0.192
Gretel	0.880	0.885	0.758	0.663	0.804
TAB	0.867	0.888	0.749	0.685	0.490
ai4privacy	0.715	0.774	0.643	0.488	0.500
Nemotron	0.855	0.863	0.773	0.639	0.656
MAPA	0.236	0.267	0.486	0.416	0.250
Internal (40-type)	0.720	0.829	0.616	0.479	0.466
ShieldFlow	0.901	0.911	0.847	0.646	0.839
ShieldFlow-xl	0.859	0.871	0.797	0.658	0.842

• Best overall across general/multilingual benchmarks and the ShieldFlow product gate; beats SauerkrautLM-GLiNER and the LFM demo on every benchmark except MAPA's idiosyncratic date-as-date_of_birth labeling convention.
• Detection-tier (did it find the PII span, ignoring fine type — the metric that matters for redaction) is markedly higher than exact-type, e.g. Internal 0.83 / ShieldFlow 0.91.

Usage

⚠️ Loads custom code via trust_remote_code=True (the model wraps a trust_remote_code encoder).

Install the required packages:

pip install torch transformers huggingface_hub

Run PII detection:

import importlib.util
import sys

from huggingface_hub import hf_hub_download
from transformers import AutoModelForTokenClassification, AutoTokenizer

model_id = "LiquidAI/LFM2.5-Encoder-350-PII-Detector"

helper_path = hf_hub_download(model_id, "pii_hybrid_decode.py")
hf_hub_download(model_id, "context_cued.py")
sys.path.insert(0, helper_path.rsplit("/", 1)[0])

spec = importlib.util.spec_from_file_location("pii_hybrid_decode", helper_path)
hd = importlib.util.module_from_spec(spec)
spec.loader.exec_module(hd)

tok = AutoTokenizer.from_pretrained(model_id, trust_remote_code=True)
model = AutoModelForTokenClassification.from_pretrained(model_id, trust_remote_code=True).eval()

spans = hd.predict("Email Dr. Laura Schmidt at laura@charite.de.", tok, model)
print(spans)

📬 Contact

• Got questions or want to connect? Join our Discord community
• If you are interested in custom solutions with edge deployment, please contact our sales team.

Citation

@article{liquidAI2026Encoders,
  author = {Liquid AI},
  title = {LFM2.5-Encoders: Fast at Long Context, Even on CPU},
  journal = {Liquid AI Blog},
  year = {2026},
  note = {www.liquid.ai/blog/lfm2-5-encoders},
}