Is this possible to run on Windows?
I had problems getting it to run on Windows 11 because of Triton.
I have to say I am not really a python developer and know the bare minimum.
Brother i recommend you to use Linux / UNIX - WSL2 if you wanna use on windows only - Cause i would say - NVIDIA developed LLM's work's the best on Linux /UNIX
Alternatively you could have set environment variable TORCHDYNAMO_DISABLE=1 for the docker container. This will make the pytorch runtime less efficient. Speech still seemed very snappy, I didn't notice any large performance hits.
@SirDaveWolf
nvm i fixed that bug. if you are not using ssl then run the browser with localhost:port , if you are using ip then you need ssl. just create a temp folder and do this instead :
python -m moshi.server --ssl path/to/temp/folder
@SirDaveWolf nvm i fixed that bug. if you are not using ssl then run the browser with localhost:port , if you are using ip then you need ssl. just create a temp folder and do this instead :
python -m moshi.server --ssl path/to/temp/folder
Yes, I fixed it by running with SSL. Good to know that localhost:port works without SSL! Ty!
Hey guys, I'm trying to run on windows 11 as well. It runs but when I start it the latency climbs up about a second every second, and the audio is stuttery. after 10 second the latency is like 10 seconds long. Anyone know what's going on?
