Some suggestions that I think would work well:
Hey Matey! (Or just ‘Hey Mate!’, if that could work)
Hey Homie! (Since this is Home Assistant, after all)
Marvin! (For the Hitchhiker’s Guide fans)
Bishop! (Aliens)
Cerebro! (X-Men)
Ive got it the .yaml is wrong configured
The original was:
{
"type": "micro",
"wake_word": "Okay Computer",
"author": "Michael Hansen",
"website": "https://www.home-assistant.io",
"model": "okay_computer.tflite",
"trained_languages": [
"en"
],
"version": 2,
"micro": {
"probability_cutoff": 0.97,
"feature_step_size": 10,
"sliding_window_size": 5,
"tensor_arena_size": 30000,
"minimum_esphome_version": "2024.7.0"
}
}
and i changed it to:
{
"type": "micro",
"wake_word": "computer",
"author": "Leland Olney",
"website": "https://github.com/JohnnyPrimus/Custom_V2_MicroWakeWords",
"model": "computer.tflite",
"trained_languages": [
"en"
],
"version": 2,
"micro": {
"probability_cutoff": 0.66,
"feature_step_size": 10,
"sliding_window_size": 10,
"tensor_arena_size": 22860,
"minimum_esphome_version": "2024.7.0"
}
}
also you need to use the RAW link to configure it.
micro_wake_word:
models:
- model: "https://raw.githubusercontent.com/oOJoshOo/Custom_V2_MicroWakeWords/refs/heads/main/models/computer/computer.json"
EDIT: It installs then, but the model doesnt load and crashes the ESP after that
This is excellent, would you consider “hello computer”. A great not to Star Trek 4.
The computer one doesn’t work.
Sorry to “resurrect” this thread, but maybe you have an opinion/answer: if i only want a wake word to work for me anyway, and if i had the time to do so: wouldn’t it then produce better results if i recorded myself 1000 times over the stretch of a few days in my own apartment? Different times of day, different positions, different devices, different moods, etcetc
Your samples will be augmented into more by cahnging pitch and rirs and should work Ok its when you have 1000 initial samples that are not your voice that are this American English with very little variation it becomes overfitted to that American English voice if you want to overfit to your voice, yeah then do so and similary it will not work very well for anyone else unless equally similar, but for you if there augmentation and training is any good it should work quite well for you.
I did some models a long time back using my voice and fewer samples and they worked well.
ProjectEars/dataset/reader at main · StuartIanNaylor/ProjectEars · GitHub was a cli reader that collected wake word but I was using the same mic for sample collection and KWS.
A Wakeword model is just a dumb image classification model statistically comparing spectragraph/mfcc images to the ones being created by your voice throught the preprocessing algs in place on the device you use.
So the best samples are from the device of use as every mic device and alg has a signature that will just increase entropy and make things less accurate.
Also distance its pointless recording farfield samples as you can only detect RiRs (Room impulse response aka reverberation) by having multiple mics that will be able to detect the TDOA (Time difference of arrival). In a mono sample as sound bounces off all surfaces the further distances mix into the sample at longer times and reverberation at a mono source just becomes difference in the spectrograph again increasing entropy and lowering accuracy.
So just record broadcast like mic differences as RiRs are usually augmented afterwards.
Coco please! or Hey Coco
I always wanted these
Jarvis
Alice
Jinx
R2
Janice
Travis
Of course without the Hey’s or OK’s
Hey @Pold,
Did you find a solution for this ?
I ran into the same issue.
My .tflite model works fine on the OpenWakeword Addon in HA, but as soon as I select it on the VPE, it crashes instantly without any signs in the logs.
With a quick search I found lots of models trained that we can use for micro wake word.
Hope this helps anyone 
There is a dearth of options for spanish speaking wakewords. Would love one since now the VPE can run two wake words at once.
Does anyone know a simple way to train this?
I tried to run this notebook but it crashes on the second step:
"ModuleNotFoundError: No module named 'piper'"
"ValueError: rate must be specified when data is a numpy array or list of audio samples."
Are you still offering this service? I’ve been hitting a wall for three days straight trying to train a couple wake words. Colob is broken, my mac is too old and it can’t do it in my pi. If someone can train a couple wake words for me i would be greatly appreciative.
You can do it yourself. See this repo: GitHub - TaterTotterson/microWakeWords: Collection of microWakeWords · GitHub
So I have a couple of wake words that I’m using from the TaterTotterson repos mentioned above…
I have them working on my atom echo for on-device wake word. It’s great! I had to get the full yaml file in the esphome device builder instead of using the packages that link to the github.
The code added looks like this:
models:
- model: /config/www/assistant_models/smithers/smithers.json
id: Smithers
I uploaded the .json and .tflite files into my /config/www/ folder to access the model.
I’ve tried to do the same for my VPE’s but just can’t get the configuration to work because it’s much more complicated code and I’m a dingus apparently. I copied the full VPE yaml into the device builder, added the custom model and installed it. It installed, but then it wouldn’t run because it couldn’t resolve the IP address (ie - wasn’t connecting to wifi anymore)
Can anyone help me with where/how I can add this same type of custom wakeword model to the VPE yaml or use the packages, but just add my own wakeword model?
Appreciate’cha!
Okay I am silly. I went way too deep on trying to do this on my own instead of doing a better search here in the forums.
There is another post here that has working yaml:
I’ll see myself out.
Hi, I don't know if it's still possible to request it, but I've always wanted to have "Oye Sirka" with Peninsular Spanish pronunciation. It would be great. Thank you very much.
Does the microwakeword creates the offline models for the esp32 s3 (respeaker lite).
yes that is correct