Skip to content
Other

Kerb Plate Nano

on-device license plate detector for mobile apps

2 min readupdated

Kerb Plate Nano is built to detect license plates right on a smartphone so they can be blurred before any frame leaves the device. It is a detector rather than an OCR (optical character recognition) model: it finds the bounding box coordinates of license plates without reading the text on them. Built on the RF-DETR Nano architecture, it was trained on a diverse multi-region dataset and exported to ExecuTorch (.pte) format for React Native apps. It is a solid fit for mobile developers who care about user privacy and on-device video processing.

Key features

  • Accurate localization: detects a single plate object class in images.
  • Mobile optimization: exported to ExecuTorch with the XNNPACK backend for efficient on-device performance.
  • Privacy protection: lets you blur license plates directly on the phone before uploading anything.
  • Low detection threshold: tuned for high recall to prevent privacy leaks.

What you can do

  • Privacy-first video: integrate it into a mobile app to automatically blur license plates in photos or video streams.
  • React Native development: drop the ready-to-use model into your app via the react-native-executorch library for local inference.

Example

const det = await ObjectDetectionModule.fromCustomModel(ptePath, {
  labelMap: { PLATE: 0 },
  preprocessorConfig: {
    normMean: [0.485, 0.456, 0.406],
    normStd: [0.229, 0.224, 0.225],
  },
  defaultDetectionThreshold: 0.3,
});
Read moreDetails
Overall mAP50 is 0.8014 on the validation split.
Trained on 13,126 images using seed 20261116.
Requires compliance with upstream dataset license obligations when redistributing.
Discuss Telegram 1 11

Will it run?

Runs in BF16

Fits without quantisation.

  • BF16weights 0.1 GB · ready files0.6 GB
  • FP8 / INT8recommendedweights 0.1 GB0.6 GB
  • 4-bitweights 0.0 GB0.5 GB

Tip: go 8-bit.FP8 / INT8 is within about a percent of full precision in quality, takes half the memory and runs faster on RTX 40/50 and H100. Full precision only pays off for fine-tuning.

What to get for this model

Massed ComputeOpen Massed Compute

Your own GPU: FP8 / INT8 — from RTX 2060 · BF16 — from RTX 2060

Massed Compute hourly prices, November 2026. The discounted price applies with the coupon at checkout.

Estimate: weights plus working memory. Long videos, high resolutions and long contexts need more.

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.

    # Shaq2/kerb-plate-nano
    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 Shaq2/kerb-plate-nano --local-dir ~/models/kerb-plate-nano
    # The model is in ~/models. How to run it is in the model card (link above).
    Add more LoRAs to kerb-plate-nano
  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 kerb-plate-nano (quants, finetunes, LoRAs) will show up under the bell.

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

About the model

Shaq2/kerb-plate-nano
Author
Shaq2
Category
Other
Released
Posted

On NeyroLearn

Hugging Face

111

Ask AI about this model

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

Comments