misyaguziya
[Issue]: Bad software performing.
Debug Tips
- [X] I'm sure I've read this project's Issues of README.
What happened?
The VRCT seem to be performing not really well. My settings are on a picture. It seems that VRCT cannot pick any lines and transcribe them for some reason. When i switch to base model it works 50/50, but either way the transcribing is very faulty and barely helps with supporting a conversation. Or, it doesn't work at all, because for some reason it doesn't catch the sound, or at least it seems so because there is no output.
I am pretty sure the case is that my CPU isn't doing too good handling this. I have Ryzen 5 5600x
Very waiting for a new update!!
APP Version
2.2.9
Runtime Environment
Windows 10
Country/Region
Europe
Relevant log output
No response
Screenshots
No response
Code of Conduct
- [X] I agree to follow this project's Code of Conduct
If the weight data is large, processing is required for that amount of data, and if it is not matched to the CPU specifications, the transcription results may not come back at all. We speculate that this may be due to the weight data of the AI model being too large for the CPU specifications to process. Why not try lowering the model to a ‘tiny model’ or some other model with a smaller capacity?
If the weight data is large, processing is required for that amount of data, and if it is not matched to the CPU specifications, the transcription results may not come back at all. We speculate that this may be due to the weight data of the AI model being too large for the CPU specifications to process. Why not try lowering the model to a ‘tiny model’ or some other model with a smaller capacity?
I've tried it with smaller model, it is indeed has higher chances of working, but it still picks up voice unpredictably...
This seems to be a device performance issue, so it may not be possible to resolve it. If you have any further questions, please open this ISSUE.