-
Notifications
You must be signed in to change notification settings - Fork 232
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Raspberry Pi 4 Buster Not Working with Binary Installation of Mycroft-Precise #90
Comments
Hello, |
Hey @tdy91, I looked at your write-up - great work there! I will try out some of your fixes. Regarding that particular error:
How did you solve it? Do you know where the problem was coming from? Yours, |
Have you attempted to use the preconfigured Buster/Mycroft image for Raspberry Pi? I'm curious if it shows any problem. The most recent (aka "unstable") Picroft image is built on Buster Lite. See: |
Hey @TheCedarPince, Regarding 'src/hostapi/alsa/pa_linux_alsa.c', line: 924 specific error, i don't try to solve it because in my tests it does not stop working. Hey @penrods, I don't try Buster/Mycroft image for Raspberry Pi. Just spent few hours to make working what i explain in issue #94 ! Yours, |
Alright - update @tdy91 I have not gotten a chance to try all your different suggested methods but am now getting this error:
I have not gotten it figured out just yet but when I have more time I will make sure to go through your guide. The thing I am trying to do is avoid using a virtual-environment but that may not be possible... I will keep you posted! Any thoughts so far? Yours, |
Hey @TheCedarPince, I don't understand what you mean saying 'I am trying to do is avoid using a virtual-environment but that may not be possible...' because I don't know exactly your installation (which virtual environment, which relationship with raspberry pi 4 ?). For your information, while my tests I do everything on the Raspberry Pi 4 + Buster lite (installation and model creation), nothing on another machine or virtual machine. In your last error log, i don't understand too where, 'precise/scripts/stream.py' come from, because following my tests I have no stream.py file, either precise/scripts/engine.py in case of 'source install', either precise-engine_0.3.0_armv7l/precise-engine in case of 'binary install'. But it does probably not matter. This said, I had this kind of error ending with 'ValueError: could not convert string to float:' when precise engine script does'nt find or understand model *.pb or *.pb.params for instance. In your case, I note that you seems to create your model on an Ubuntu Linux Machine (maybe the virtual environnment you are talking about) and i note also at the end ogf your log error From your error message, to verify this hypothesis and qualify your raspberry pi 4 installation, i think you can just download hey-mycroft.tar.gz from precise-data repository https://github.com/MycroftAI/precise-data/tree/models/hey-mycroft.tar.gz If it works, you will just to have train your own model on the raspberry pi 4 itself or solve the problem otherwise. Yours, |
Hey @tdy91 ! I went back through your instructions and I actually got things working! I am not going to close this issue yet as I discovered a lot of fixes that need to be incorporated into the documentation. I am going to do a write up for anyone else who has the same issues. There were some things in your instructions that were missing so in the write-up here, I will include the missing steps. Moreover, @krisgesling I should have this ready for reference in MycroftAI/documentation#139. Will let you know when this is ready to go! Thanks all! |
Useful Info
OS: Raspbian Buster
Python: Python3.7
Mycroft-Precise Version: 0.3.0
Issue:
Hello Mycroft Precise Team (e.g. @MatthewScholefield)!
I have been trying to get precise-runner working with my Raspberry Pi 4 running Raspbian Buster.
After creating my own model on a Ubuntu Linux machine, I tried to get it Mycroft-precise to use it on the Pi (here is my model and its associated params. If anyone wants to test out the model, the wake word is "Automaton". I followed the steps to train my model and it yielded a 98.5% accuracy so I feel like it is pretty robust.)
Once I transferred everything over to my Pi, I ran into a solid wall where I spent upwards of three hours debugging things to no avail. I ended up trying to build from source as well but still no luck.
Here is the error I consistently get:
And this was after hours of tampering with my set-up. What I think is happening is something is happening with PyAudio but I know not what. I tried this also on Python3.5 on my Ubuntu Linux Machine and I am getting the same error.
Any thoughts about how to work around this? I would appreciate any thoughts here as I very much wish to use this software! Let me know if I can provide any more assistance
Steps to Reproduce
Follow Source Install instructions
Create Samply Python Program
sudo python3 example.py
with my model now in place as well as executing this from the virtual environment.Additional Info:
Here is my latest .asoundrc file that I created in my home directory:
.asoundrc
Here is my alsa.conf file in my
/usr/share/alsa/
directory:alsa.conf
Here is the output of
arecord -l
:Here is the output of my Python virtual environment:
The text was updated successfully, but these errors were encountered: