Hi there,
I’m trying to set up kaldi running on my RPI3B+ setup, on an FR profile.
I’ve read some thread here about enhances fr profile, but I’m currently not able to download base files.
In original set up I just can’t enable kaldi :
Do speech recognition with kaldi on this device
Not compatible with this profile
So I forced Kaldi in profile.json by adding :
"system":"kaldi",
In “speech-to-text” setting
After rhasspy restart it ask :
Some files must be downloaded for your selected profile (fr).
Rhasspy will not work correctly until these files are downloaded.
“missing”:
“/profiles/fr/kaldi”:
0: “speech_to_text.system”
1: “kaldi”
That’s sound great! that"s the folder I miss to dl FR enhanced files.
=> But few seconds after hitting download button -> error :
Compressed file ended before the end-of-stream marker was reached
I do not know what I can do with https://github.com/synesthesiam/rhasspy-profiles/releases/tag/v1.0-fr
So I’ve tried to copy kaldi-generic-fr-tdnn_250-r20190930.tar.xz to a kaldi folder in my profile and point to that folder in the json, but event if it succeed in training, it doesn’t work at all.
What could I do?
[note] My setup in my profile!
Many thanks
Ben