FrontierDK
FrontierDK
Well, I'm also training from scratch with my own models, and KenLM does seem to improve the recognized text a lot (far less errors) - but being installed in VMs,...
> I was thinking the same about emotions and or expressions. I don't know how it will be handled in Piper. An easy way, not requiring any metatags or SSML...
If I had to guess, it would take a lot of new training data to be able to do this. If in a rush, try looking at these: **Commercial solutions**...
L4cache, try asking for a full shared x86 build, including libfdk, libjxl etc.
> Running the same configuration (mabs.ini, ffmpeg_options and mpv_options) resulted in the build passing for me. What is the folder path mabs is in? Perhaps your base path is too...
Just updated to the latest on both my x86 and x64 VMs (both run identical Windows 10 x64), one has x64 building of FFMPEG, one has x86. Last night, the...
To me not having trained ANY dataset using Piper yet, it sounds to me like you're running out of VRAM?
It can, but it's slower, at least for single inference. I got it working on Ubuntu (in a virtual PC with access to a RTX 3060 via PCI-E passthrough). If...
I'm a newbie and I haven't even tried this one yet, but isn't this a voice cloning tool? I.e. you find a friend who will let you record a few...