Skip to content
Other

LumoHub Physics-Feature Event MLP

a compact motion recognition model for ESP32

2 min readupdated

This physics-informed Multi-Layer Perceptron (a feedforward neural network) is a compact model built to run on the LumoHub microcontroller (ESP32-S3). It processes 16 biomechanical features extracted from 3-second windows of 6-axis IMU (inertial measurement unit) data, such as free-fall depth, impact pulse width, and gravity tilt angle. The model outputs high-accuracy event classifications, including fall detection. It has a tiny footprint—just 3,365 float values taking up 13,460 bytes of Flash—and avoids false positives when the device itself is dropped or bumped on the wrist. It is a solid fit for embedded systems and wearables developers building real-time motion analysis.

Key features

  • 99.73% accuracy and 99.73% Macro F1 on a held-out test set (N = 4,500)
  • Zero false alarms (0.00%) from wrist impacts and device drops
  • Ultra-compact footprint: 3,365 float values and 3,232 MACs/inference
  • Resilient to BLE packet loss and sensor noise

What you can do

  • Integrate it into ESP32-S3 firmware using the exported C99 weights from the header file
  • Build reliable fall detection for wearable medical devices or fitness trackers
  • Analyze physical motion parameters (accelerations, jerks, tilt angles) directly on microcontrollers
Read moreDetails
Architecture: Input 16 -> Linear(16->64) + ReLU -> Linear(64->32) + ReLU -> Linear(32->5) with batch normalization weights.
Benchmarks: Binary fall detection sensitivity (classes 3 and 4) reaches 99.94%. Under 10% BLE packet loss, accuracy drops to 96.56%.
Integration: Weights and scaling constants are exported to lumo_hub_feature_model.h (lumo_imu_ai.c).
Discuss Telegram 0 28

Will it run?

Runs in BF16

Fits without quantisation.

  • BF16weights 0.0 GB · ready files0.5 GB
  • FP8 / INT8recommendedweights 0.0 GB0.5 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.

    # fragilov/lumohub-imu-event-mlp
    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 fragilov/lumohub-imu-event-mlp --local-dir ~/models/lumohub-imu-event-mlp
    # The model is in ~/models. How to run it is in the model card (link above).
    Add more LoRAs to lumohub-imu-event-mlp
  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 lumohub-imu-event-mlp (quants, finetunes, LoRAs) will show up under the bell.

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

About the model

fragilov/lumohub-imu-event-mlp
Author
fragilov
Category
Other
Released
Posted

On NeyroLearn

Hugging Face

280

Ask AI about this model

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

Comments