It would be useful to mention in the README to submit not only the .tflite file but also a note with the params and the evaluation of each submitted model. Maybe even reject PRs that do not provide them.

It's very easy to provide these when submitting. As the number of models increases it will be hard to understand which are high quality and which are not.

I keep getting: ValueError: Model provided has model identifier 'yloa', should be 'TFL3'
When I try to load some of the models like Alice, Hey_Honey, Hey_zelda, but not alexa_v0.1

I am using openwakeword in a container and loaded it with the --custom-model-dir set to /custom. The wake word shows in home assistant. Running a test for activation, everything fails though. I get the following error in openwakeword.

ERROR:root:Unexpected error in wake word thread (echoh)
Traceback (most recent call last):
  File "/usr/src/.venv/lib/python3.11/site-packages/wyoming_openwakeword/openwakeword.py", line 248, in ww_proc
    ww_model = tflite.Interpreter(model_path=str(ww_model_path), num_threads=1)
               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/src/.venv/lib/python3.11/site-packages/tflite_runtime/interpreter.py", line 464, in __init__
    self._interpreter = _interpreter_wrapper.CreateWrapperFromFile(
                        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
ValueError: No subgraph in the model.

INFO: Created TensorFlow Lite XNNPACK delegate for CPU.

Comment your wanted wake-words below this comment.