Hello, i have a few questions regarding install Larynx in NanoM1+ and OrangePiZeroLTS. Before i get into my questions, i really love Rhasspy and Larynx. I managed to install Rhasspy (via docker) in my NanoM1+, and Larynx sounds great. Then i managed to install Larynx itself in my Virtualbox (in my Macbook). Now i wanna take a further step, is to experiment with 2 of my armv7l devices. Here are my questions:
-
It seems that to install Larynx, it requires to have Pytorch installed, but i cannot find any official Pytorch wheel for armv7l. I did found “3rd party” one which compatible on Raspi4, so i tried it but i stuck on installing the following packages : Scipy and Scikit-learn. (Please refer my screenshot below).
If so, is there official Pytorch wheel for armv7l devices ? -
Whenever i tried to to run “pip3 install -r requirements.txt”, my device just went hang, 100% RAM usage. I have to increase the tmpfs (tmp) size to 1GB since some error said my devices doesn’t have enough storage. Both of my devices have 512MB RAM. My question, what are the minimum hardware requirements to run Larynx ?
-
Regarding the first question, as i cannot install Scipy and Scikit-learn properly (errors are like cannot build the wheel on this devices …) and it does take hours when installing/building the wheels but in the end return errors. Is it normal that it take hours or it is just my hardware is not strong enough ?
-
If i managed to install Rhasspy (docker installation) in my NanoM1+ without any problems, how come installing Larynx alone does pose a bit of problem? Is there any pre-requirements in Rhasspy that i need to install beforehand which can make Larynx installation smoother ? (sorry for this question, i’m no expert in Docker)
Any helps, opinions are welcome. Correct me if i made mistake. I really hope that i can install Larynx in one of my devices.
p/s : both my devices running armbian OS.
Thank you in advance.