Overview: Support for Docker image BirdNET-Go

Application Name: BirdNET-Go
Application URL: https://github.com/tphakala/birdnet-go

Documentation: https://github.com/tphakala/birdnet-go/wiki

 

Make sure to look at the complete documentation

 

Please post any questions/issues relating to this docker template you have in this thread.

Note: this is not a support thread for BirdNET-Go

A reminder that you may need to install the following plugin, if you are using a USB microphone or a sound card, rather than a camera stream.

https://raw.githubusercontent.com/ich777/unraid-sound-driver/refs/heads/master/sound-driver.plg

thanks @u03mdr
but i doubt (even if it is fairly possible) they will install a usb microphone on the unraid server.....
but that's a great reminder.

i will add it in the readme

Hi, thank you for this wonder app! .Question: When I run the command "birdnet-go benchmark" the TensorFlow Lite XNNPACK model is way slower than the standard model.

Standard 48.0 ms 20.50 inferences/sec

XNNPACK 609.0 ms 1.70 inferences/sec

Is that because I’m running it on unraid with an AMD Ryzen 5 3600 instead of like a raspberry pi? Would it do better with the regular TensorFlow model instead of Lite version? At any rate, until that happens should I change the "usexnnpack" variable in the config.yaml to false? I'm having it listen to two rtsp steams from two ip cameras and the container is consistently using about 10% of my processing power. Just thought that was a little much. Thanks again for this get container in the unraid community apps store!!!

Edited September 4, 2025Sep 4 by Pharmhero

Hello,

Thanks for the compliments, but it is not my app....
but i am totally agree it is a great app :)

your question is a excellent candidate to ask in their github : https://github.com/tphakala/birdnet-go
you will have more expert answer.

Thanks

Got it. Thank you for the quick reply! And I still want to thank you for bringing this to the Unraid Community Apps store where I learned about the BirdNET project. I wrongly assumed you had something to do with the packaging of the container.

no worries,
i really enjoy this app.