Skip to content
Other

PII Image Redactor

находит приватные данные на скриншотах

2 min read

Модель от screenpipe для детекции PII-регионов прямо на изображениях: скриншотах, кадрах computer-use и интерфейсах приложений, которые видит AI-агент. Это ONNX-файл примерно на 109 МБ, который возвращает пиксельные bbox для 12 категорий приватных данных.

Ключевые особенности

  • Один и тот же файл работает на macOS, Windows и Linux через CoreML, DirectML, CUDA или CPU; провайдер выбирается при загрузке, вендорные GPU SDK не нужны.
  • Детектит 12 категорий: private_person, private_email, private_phone, private_address, private_url, private_company, private_repo, private_handle, private_channel, private_id, private_date и secret — пароли, API-ключи, JWT и т.п.
  • На ScreenLeak при IoU ≥ 0.30: region zero-leak 98.9%, oversmash 0.0%.
  • Около 120 мс p50 на Apple Silicon через CoreML.
  • Для редакции советуют сплошной чёрный прямоугольник, а не блюр: блюр можно обратить super-resolution-атакой.
  • Лицензия CC BY-NC 4.0 — только некоммерческое использование; для коммерции указан контакт louis@screenpi.pe.

Пример

import numpy as np, onnxruntime as ort
from PIL import Image
sess = ort.InferenceSession('rfdetr_v11.onnx', providers=['CPUExecutionProvider'])
img = Image.open('screenshot.png').convert('RGB')
arr = np.asarray(img.resize((512, 512), Image.BILINEAR), np.float32) / 255.0
arr = ((arr - [0.485,0.456,0.406]) / [0.229,0.224,0.225]).transpose(2,0,1)[None]
boxes, logits = sess.run(None, {sess.get_inputs()[0].name: arr.astype(np.float32)})
# boxes/logits → фильтр по score и закраска bbox
ПодробнееБенчмарк: 98.9% region zero-leak и 0.0% oversmash на held-out split ScreenLeak при IoU ≥ 0.30; это верхняя граница для условий обучения, а не гарантия на реальных экранах.
Класс региона ненадёжен вне распределения — закрашивайте все найденные bbox, не фильтруйте по label.
Обучение только на синтетике; оценка на английском/латинице, CJK, арабика и кириллица не проверялись.
Не security boundary: противник может подобрать layout для обхода.
Discuss Telegram 0

Will it run?

Working out the memory needs

The bot is still collecting this model's weight sizes; the estimate will appear here in a few minutes. Meanwhile you can rent a GPU with headroom.

Massed ComputeOpen Massed Compute

Run it on Massed Compute

The script downloads the model to the rented machine. How to run it is in the model card.

  1. 1. Rent a machine with a fitting GPU

    Open Massed Compute
  2. 2. Paste the script into the machine's terminal

    Connect over SSH (the Massed Compute panel shows the command) or open a terminal on the machine's desktop.

    HF_TOKEN / CIVITAI_TOKENThe script starts with token lines: paste yours between the quotes. The Hugging Face token is for gated models, the Civitai one for Civitai models.

    # MylesMCook/pii-image-redactor
    mkdir -p ~/models
    curl -LsSf https://astral.sh/uv/install.sh | sh
    source "$HOME/.local/bin/env"
    uv tool install "huggingface_hub[cli]"
    hf download MylesMCook/pii-image-redactor --local-dir ~/models/pii-image-redactor
    # The model is in ~/models. How to run it is in the model card (link above).
    Add more LoRAs to pii-image-redactor
  3. 3. Use it

    The model is in ~/models. How to run it is in the model card (link above).

Stop the machine in the panel when you are done: it bills by the minute while it runs.

How to speed it up

Picked for this model type. Choose your hardware above to see what fits you exactly.

  • FlashAttention 44.0.0b40from November 18, 2026new release

    fastest exact attention on B200

    Kernels for Blackwell (beta). Exact attention without quantisation — fine for everything, LLMs included.

    How to enable: pip install flash-attn-4; vLLM and SGLang pick it up themselves.

    Earlier generations (2)
    • FlashAttention 3 — The Hopper (H100/H200) version with FP8.
    • FlashAttention 2 — The standard for RTX 30/40/50 and A100: exact attention without the memory cost of long contexts.
  • torch.compile2.15.0from October 28, 2026new release

    10–30% for free

    Compiles the model for your GPU. Slower first run, faster afterwards.

    How to enable: model = torch.compile(model); in ComfyUI the TorchCompileModel node.

Model family

New posts about versions of pii-image-redactor (quants, finetunes, LoRAs) will show up under the bell.

No known relatives of this model yet. All derivatives on Hugging Face

About the model

MylesMCook/pii-image-redactor
Author
MylesMCook
Category
Other
Released
Posted

On NeyroLearn

Hugging Face

00

Ask AI about this model

Answered by Gemini 4 · the chat is kept only in this tab

Comments