Documentation: Start from Commandline/Ubuntu

[phigjm]

Is there some documentation about how to start the project from the command line?

I would like to try to run the project from Linux and I also want to modify the 3s segment size.

What I did so far:

Install python 3.11 (the ubuntu default version 3.12 doesn't work).

I followed this guide.

Than I create a venv with the 3.11 version and activated this.

This allowed me to install the package with:

pip3 install . called within the cloned BirdNet-Analyzed root folder.

Sadly, there seem to be some pip requirements missing:

matplotlib
gradio
plotly
tensorrt
keras_tuner 

and the UI itself also need some additional apt packages: eighter (what worked for me):

pip install pywebview[qt]
sudo apt install python3-pyqt5 python3-pyqt5.qtwebengine python3-pyqt5.qtwebchannel libqt5webkit5-dev

or (not tested):

pip install pywebview[gtk]
sudo apt install python3-gi python3-gi-cairo gir1.2-gtk-3.0 gir1.2-webkit2-4.1

to run the project i used:

python3 -m birdnet_analyzer.gui

Now The UI is working but sadly it seems to be that the training is already stuck in the loading process...

W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT
Loading training data...
-loading "firstLabel": 0%

It would be great if you have some further documentation or hints for me.

I am also willing to write this down into a clean documentation if it is really missing.

[yurish]

I am not adding another issue, but stumbled on similar issue. keras-tuner is mentioned as an optional dependency in the [train] set, but apparently it is being used in model.py. That's why running analyze module or in GUI (after pressing Analyze button) you will get ModuleNotFound exception. Shouldn't it be included in the common dependencies of the project then?