Hi,
I recently bought a Syntiant TinyML board and started exploring the edge impulse ecosystem.
Unfortunately, I couldn’t find any documentation on the posterior parameters in the deployment section.
Can anyone give me some more information on the parameters?
{
"states": [
{
"timeout": 0,
"classes": [
{
"label": "go",
"phwin": 5,
"phth": 0.95,
"phbackoff": 20,
"phaction": 0,
"phaction_arg": 0,
"smoothing_queue_size": 1
}
]
}
],
"quantization_arguments": [
null,
[
"uniform",
4,
"min"
],
null,
[
"uniform",
4,
"min"
],
null,
[
"uniform",
4,
"min"
],
null,
[
"uniform",
4,
"max"
],
null
],
"audio_pdm_gain": 12
}
I am especially interested in the phwin, phth, parameters etc. As well as the quantization_arguments and audio_pdm_gain.
Cheers,
Matt