privacy-filter

What it does

This model performs token classification to identify personal identifiers such as account numbers, private addresses, dates, emails, person names, phone numbers, URLs, and secrets in English text. It runs locally on Apple Silicon with MLX, using 8-bit quantization, and the weights file is about 1.39 GiB.

Best for

• Redacting personal data from clinical notes
• Finding identifiers in research documents
• Local privacy checks in medical apps

What it detects (8 labels)

• Account number
• Private address
• Private date
• Private email
• Private person
• Private phone
• Private url
• Secret

Quality

No evaluation score is published for this model.

Use it

pip install openmed

from openmed import analyze_text
result = analyze_text("Patient prescribed 10 mg aspirin for hypertension.", model_name="OpenMed/privacy-filter-mlx-8bit")
for e in result.entities:
    print(e.label, e.text, round(e.confidence, 3))

pip install "openmed[mlx]"

from openmed import analyze_text
from openmed.core.config import OpenMedConfig
result = analyze_text("Patient started imatinib in March.", model_name="OpenMed/privacy-filter-mlx-8bit", config=OpenMedConfig(backend="mlx"))
for e in result.entities:
    print(e.label, e.text)