Token Classification
Transformers
Safetensors
lfm2
liquid
lfm2.5
bidirectional
masked-lm
encoder
pii
ner
privacy
multilingual
custom_code
Instructions to use LiquidAI/LFM2.5-Encoder-350M-PII-Detector with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use LiquidAI/LFM2.5-Encoder-350M-PII-Detector with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("token-classification", model="LiquidAI/LFM2.5-Encoder-350M-PII-Detector", trust_remote_code=True)# pip install -U transformers accelerate # Load model directly from transformers import AutoTokenizer, AutoModelForTokenClassification tokenizer = AutoTokenizer.from_pretrained("LiquidAI/LFM2.5-Encoder-350M-PII-Detector", trust_remote_code=True) model = AutoModelForTokenClassification.from_pretrained("LiquidAI/LFM2.5-Encoder-350M-PII-Detector", trust_remote_code=True, device_map="auto") - Notebooks
- Google Colab
- Kaggle
Add entity-type list (40 types x 11 domains) to model card
Browse files
README.md
CHANGED
|
@@ -18,6 +18,24 @@ byte-BPE boundary fragmentation.
|
|
| 18 |
locale-personas × scenarios × cue/inline/structured embedding, LLM-judge-filtered,
|
| 19 |
contamination-cleaned against all eval sets).
|
| 20 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 21 |
## Benchmarks (18-locale-filtered, partial-F1, hybrid decode)
|
| 22 |
|
| 23 |
Comparison vs the prior release (v8) and external models (SauerkrautLM-LFM2.5-GLiNER, the
|
|
|
|
| 18 |
locale-personas × scenarios × cue/inline/structured embedding, LLM-judge-filtered,
|
| 19 |
contamination-cleaned against all eval sets).
|
| 20 |
|
| 21 |
+
## Entity types (40 PII types across 11 domains)
|
| 22 |
+
|
| 23 |
+
The model emits canonical `type` labels (BIOES token classification + an inference-time hybrid regex decode for format-bound types). Full label set:
|
| 24 |
+
|
| 25 |
+
| Domain | Types |
|
| 26 |
+
|---|---|
|
| 27 |
+
| **Identity** | `identity.person_name`, `identity.ssn`, `identity.national_id`, `identity.passport`, `identity.drivers_license`, `identity.date_of_birth`, `identity.tax_id` |
|
| 28 |
+
| **Contact** | `contact.email`, `contact.phone`, `contact.address`, `contact.postal_code`, `contact.ip_address` |
|
| 29 |
+
| **Financial** | `financial.credit_card`, `financial.iban`, `financial.bank_account`, `financial.swift_bic`, `financial.crypto_wallet`, `financial.amount` |
|
| 30 |
+
| **Credentials** | `credential.api_key`, `credential.password`, `credential.private_key`, `credential.jwt`, `credential.connection_string`, `developer.login_credentials` |
|
| 31 |
+
| **Online** | `online.username`, `online.url` |
|
| 32 |
+
| **Device** | `device.mac_address`, `device.imei`, `developer.device_id` |
|
| 33 |
+
| **Location** | `location.gps_coordinates` |
|
| 34 |
+
| **Healthcare** | `healthcare.medical_record`, `healthcare.condition`, `healthcare.medication`, `healthcare.health_plan_id` |
|
| 35 |
+
| **Organization** | `org.company_name` |
|
| 36 |
+
| **Special-category** | `special.religion`, `special.political`, `special.orientation`, `special.health_status` |
|
| 37 |
+
| **Legal** | `legal.case_number` |
|
| 38 |
+
|
| 39 |
## Benchmarks (18-locale-filtered, partial-F1, hybrid decode)
|
| 40 |
|
| 41 |
Comparison vs the prior release (v8) and external models (SauerkrautLM-LFM2.5-GLiNER, the
|