Description

WekiTrainer is a Max4Live device that assists in the model training and output mapping process of Machine Learning by communicating directly with Wekinator via Ableton Live.

Version 0.7 features include:

◼︎ Mapping of • Record, 〄 Train and ▶︎ Run buttons in Wekinator directly from Ableton via keymapping, Macro or MIDI.

◼︎ 3 Continuous Outputs (Red, Green, Blue). Each output can be mapped onto any device parameter in Ableton. Their corresponding sliders can be mapped and monitored via external MIDI or Macro, helpful when training Wekinator with a new gesture. Also includes a randomizer.

◼︎ 3 Continuous Output Monitors with control of threshold, min. and max. values, output rate in miliseconds (up to 1 second) and compressing, jittering and smoothing of output signal.

❖…also, you can map “Forget Training” button, which clears all previously recorded data in the currently trained model!

Video instructions here: https://vimeo.com/296526211

If you enjoy this Max4Live device, please consider donating from 1(฿, €, $, ¥, £, …) to whatever you decide is fair to pay for its use and future development.

• Please send your donations via Paypal to

hola[at] moiseshorta.audio with the subject: WekiTrainer.

If you would like to use it for facilitating a workshop in your community about Machine Learning and Wekinator, please, feel free to reach me to my email:

hola [at] moiseshorta.audio

You can find the lite version

WekiTrainerMini

for download here:

http://www.maxforlive.com/library/device/5059/wekitrainermini

WekiTrainer and WekiTrainerMini requires the Wekinator: a free, open source software originally created in 2009 by Rebecca Fiebrink. It allows anyone to use machine learning to build new musical instruments, gestural game controllers, computer vision or computer listening systems, and more.

You can download it for FREE here:

wekinator.org

Thanks for downloading and happy WekiTraining!

###Many thanks to M4L ninja Carlo Cattano for the help debugging the device!