npm package discovery and stats viewer.

Discover Tips

  • General search

    [free text search, go nuts!]

  • Package details

    pkg:[package-name]

  • User packages

    @[username]

Sponsor

Optimize Toolset

I’ve always been into building performant and accessible sites, but lately I’ve been taking it extremely seriously. So much so that I’ve been building a tool to help me optimize and monitor the sites that I build to make sure that I’m making an attempt to offer the best experience to those who visit them. If you’re into performant, accessible and SEO friendly sites, you might like it too! You can check it out at Optimize Toolset.

About

Hi, 👋, I’m Ryan Hefner  and I built this site for me, and you! The goal of this site was to provide an easy way for me to check the stats on my npm packages, both for prioritizing issues and updates, and to give me a little kick in the pants to keep up on stuff.

As I was building it, I realized that I was actually using the tool to build the tool, and figured I might as well put this out there and hopefully others will find it to be a fast and useful way to search and browse npm packages as I have.

If you’re interested in other things I’m working on, follow me on Twitter or check out the open source projects I’ve been publishing on GitHub.

I am also working on a Twitter bot for this site to tweet the most popular, newest, random packages from npm. Please follow that account now and it will start sending out packages soon–ish.

Open Software & Tools

This site wouldn’t be possible without the immense generosity and tireless efforts from the people who make contributions to the world and share their work via open source initiatives. Thank you 🙏

© 2024 – Pkg Stats / Ryan Hefner

tflite-find-arena-size

v1.0.2

Published

Find TensorFlow Lite arena size

Downloads

13

Readme

Find TensorFlow Lite Arena Size

TensorFlow Lite for microcontrollers requires an arena size to be specified, but there's no quick way of calculating the minimum size of the arena. This repository contains an application which loads a tflite model and uses a divide and conquer algorithm to find the optimal size. The size of the arena is dependent on the target architecture, so it's advisable to run this application on the same platform as where you want to deploy the model. What we've found so far is that the arena size on a 64-bit computer (e.g. under macOS) is always bigger than on 32-bits. There's also a cross-compilation script available which builds a WebAssembly version of this application, which you can use from any Node.js or web application. It seems that the WebAssembly version approximates 32-bit embedded systems pretty well.

Example outputs

Using float32 models (non-quantized) with TensorFlow Lite 2.1.

Multi-layer perceptron with two hidden layers

| Environment | Arena size | | ------------- | ------------- | | macOS Catalina | 2272 bytes | | Node.js (WebAssembly) | 1420 bytes | | Mbed OS (Cortex-M4F) | 1392 bytes |

1D Convolutional neural net with two hidden layers

| Environment | Arena size | | ------------- | ------------- | | macOS Catalina | 26400 bytes | | Node.js (WebAssembly) | 24968 bytes | | Mbed OS (Cortex-M4F) | 24896 bytes |

Cloning this repository

This clones the repository and initializes all submodules:

$ git clone --recurse-submodules https://github.com/edgeimpulse/tflite-find-arena-size.git

Building locally (native)

  1. On Linux: Install LLVM 9 (install instructions). The default cmake version on macOS Catalina already works.

  2. Build the application:

    $ make -f Makefile.tflite
  3. Run the application and pass in a tflite file:

    $ ./build/find-arena-size some-tflite-file.tflite
    arena_size is 2264

Building locally (WebAssembly)

  1. Install Docker.

  2. Build the application:

    $ docker build -t tflite-find-arena-size .
    $ docker run -v $PWD/build:/app/build --rm tflite-find-arena-size
  3. Run the application and pass in a tflite file:

    $ node emcc/test.js some-tflite-file.tflite
    arena_size is 1420

Installing (npm)

  1. Add the package to your Node.js application via:

    $ npm install tflite-find-arena-size --save
  2. Use the package as usual:

    const FindArena = require('tflite-find-arena-size');
    const fs = require('fs');
    
    (async () => {
        let fa = new FindArena();
        await fa.init();
    
        let tflite = await fs.promises.readFile('some-tflite-file.tflite');
        let size = await fa.findArenaSize(tflite, 1 * 1024, 128 * 1024);
        console.log('arena size is', size);
    })();