Totally agree, I threw a quick filter on some public domain, actual human artwork, and cropped it square. Took me as long as it would have taken to generate that slop logo.

Here's the original source: Citation

For the love of anything even mildly good or wholesome...enough of the fucking AI slop images!

🔗 AI Logo

🔗 and probably everything else too

🔗 Including the readme

The maintainer of the project probably won't see it here, but that repo does appear to be fairly active, they've got thousands of merged PRs. If you contribute a better logo they might replace the AI one.

For the benefit of others:

Review the license!!

You may not use BirdNET-Pi to develop a commercial product!!!!

BirdNET-Pi

A realtime acoustic bird classification system for the Raspberry Pi 5, 4B, 400, 3B+, and 0W2

Icon made by Freepik from www.flaticon.com

About this fork:

I've been building on mcguirepr89's most excellent work to further update and improve BirdNET-Pi. Maybe someone will find it useful.

[...]

Introduction

BirdNET-Pi is built on the BirdNET framework by @kahst using pre-built TFLite binaries by @PINTO0309 . It is able to recognize bird sounds from a USB microphone or sound card in realtime and share its data with the rest of the world.

Check out birds from around the world

• BirdWeather

Features

• 24/7 recording and automatic identification of bird songs, chirps, and peeps using BirdNET machine learning
• Automatic extraction and cataloguing of bird clips from full-length recordings
• Tools to visualize your recorded bird data and analyze trends
• Live audio stream and spectrogram
• Automatic disk space management that periodically purges old audio files
• BirdWeather integration -- you can request a BirdWeather ID from BirdNET-Pi's "Tools" > "Settings" page
• Web interface access to all data and logs provided by Caddy
• GoTTY and GoTTY x86 Web Terminal
• Tiny File Manager
• FTP server included
• SQLite3 Database
• Adminer database maintenance
• phpSysInfo
• Apprise Notifications supporting 90+ notification platforms
• Localization supported

I use both Song Sleuth and Smart Bird ID (iOS). Fairly decent. I wouldn't bet any significant bet on it's accuracy 100% of the time. I'm no birder, but it seems to work well enough for my purposes.

It is the same back end, but this one runs continuously on a server instwad of on your phone.

I'm afraid Merlin's addition of videos of automatically running when birds are detected is the beginning of enshitification. Phone charge no longer lasts a full field day with this additional drain on battery. Bird of the day, x number of people listening, I don't need of want any of this shit. Also dislike can no longer immediately cancel recording without saving, have to wait while application buffers or something before can exit without saving, for a long recording this can take more than 15 seconds. Pity, app is generally very helpful.

I've been running birdnetpi on a raspberry pi 3 for years. It runs very well on this modest hardware. I also run birdnet-go in a docker container and use my reolink security cameras for the RSTP audio feed to that instance.

There are all sorts of AI/ML models available that can run on smaller hardware. But they are usually very specific and not general purpose, like the LLMs.

Audio processing by ai can absolutely happen on a pi.

As an example, whisper.cpp is a high performance speech recognition model, and it does a great job. The biggest version of it as a model requires less than 5GB to live fully in memory, the large model at about 1.5billion parameters, and it works great even CPU only.

The one used most often is small, which has about 244million parameters, and needs about a gig of ram and nothing more.

For a pi 5, usually the base or tiny model are used. Small could be used in most scenarios, but tiny and base are able to run real-time.

So it depends for how local - sometimes its all self contained and can do just fine on a pi, sometimes its local in that it needs an LLM endpoint to hit, but that can be running on another system entirely.

It would really depend on the project for how/where/why for a definition of local.

Specialized models can be made incredibly efficient. The big tech guys are trying to replace humans which is why their crap takes such huge servers to run. Building local-first/only just means you need to specialize the model and the pipeline to match what you're doing

I haven't verified it, but it local ML models on audio is pretty efficient and should run fine on a raspberry pi.

It can run on device on your phone https://birdnet.cornell.edu/

Listen, everything that they are doing by scorching the planet and turning the tap water brown, could be achieved on CPU or much smaller GPUs on local machines - it's a question of focused, tight training. What they're doing instead is literally attempting to build gigantic models that take in absolutely everything (burning real books in the process because fuck history we got money to make) and attempting to sell the idea that this brute force, one-thing-does-it-all approach, which requires us to give up everything else that is good in the world, is the only way forward, and yes, we must build the roko basilisk now.

It's a big, big grift.

It's a small model that runs on your machine using a TFLite model. AI/ML models come in many different sizes. Some more conducive for running locally than others. There's more info on the detection pipeline here.

I didn't write this, just found and self hosted it.