-
Hi, Maybe I'm missing something. But I start the GPTHome service. Then in the event logs I see it ticking along with "INFO:functions:Could not understand audio, waiting for a new phrase..." if I'm making random noises or speaking without using the wake word - which is completely normal and fine. And if I turn on debug I see the usual "DEBUG:openai:message='Request to OpenAI API' method=get path=https://api.openai.com/v1/models" type logs. All good. If I say something to it using the wake word "i.e. computer what is a banana" then I can see that come up in the log ok as a success, i.e. "SUCCESS:functions:Heard: "what is a banana"" - I don't see anything about a response unless I turn on debug, then I see: DEBUG:openai:message='Request to OpenAI API' method=post path=https://api.openai.com/v1/chat/completions But I don't see the actual response to my query show up as a "Success" line - unless, I restart the GPTHome service, then the response seems to come through to the logs. If I don't restart the service then no further commands seem to be picked up (at least according to the logs). Once I restart the service, I can issue one command and then no more are accepted. I know it's made it over to OpenAI as I can see the usage against the API key go up. Additionally I appear to have no audiable responses, but I don't know whether that is because something is going on with the audio with GPT-Home (speaker-test works fine, I am using a mini speaker that plugs into the 3.5 port much like the one pictured against the project), or whether I'm just not getting any response which is why I hear nothing. Any help appreciated EDIT: It's like it stops accepting commands as it never gets the answer back from OpenAI - the query reaches openAI, as I can see the API usage increment. Then no answer comes back until I stop and restart the service, the answer comes back when I do that (but I don't hear anything come out of my speaker) and I am able to issue another command. |
Beta Was this translation helpful? Give feedback.
Replies: 5 comments 18 replies
-
When Edit: Also, does this behavior repeat with other models? |
Beta Was this translation helpful? Give feedback.
-
Hi, I rebooted and still can't play sound from inside the container. asound.conf exists in the container: pcm.!default { type plug; slave.pcm "dmix0"; } Do I need to rebuild something after updating the user permissions? |
Beta Was this translation helpful? Give feedback.
-
Hi, I'm not sure if it's the same issue but when I run:
I hear no sound, there a small pause then it comes back to the prompt. When I run it with sudo, I hear the wav file.
Is this normal and does this affect the speaker output of the project? I can't hear any feedback from it yet. |
Beta Was this translation helpful? Give feedback.
-
I have exactly the opposite of what @taherafridi mentioned. I'm trying to get it work on
|
Beta Was this translation helpful? Give feedback.
-
Does pcm.!default { type hw card 3 device 0 }
ctl.!default { type hw card 3 } |
Beta Was this translation helpful? Give feedback.
Bingo - got it working. One thing in asound.conf needed changing. The line "#pcm.!default { type plug; slave.pcm "dmix0"; }" needed to be commented out, and replaced with "pcm.!default { type hw; card 0; }" - once that was done everything started working immediately!
Thank you very much for your help with this, I would have had no idea where to look at all if it wasn't for you. Out of curiosity any idea why my asound.conf had what it did for the default pcm device?