Hello,
First of all, thank you for developing Voicebox.sh. I really appreciate the work that has gone into the project and the overall voice generation experience.
For a future update, I would like to suggest adding more advanced controls to the voice generation stage. It would be very useful if users could adjust the following parameters directly from the interface:
Speed
Temperature
Top P
Top K
Repetition Penalty
GPT Conditioning
GPT Conditioning Chunk Length
Having access to these parameters would give users more control over the generated speech and allow them to fine-tune the output for different use cases.
For example, parameters such as Temperature, Top P, Top K, and Repetition Penalty could help users control the generation behavior and improve pronunciation or stability in certain cases. Speed would also be useful for adjusting the speaking rate without having to modify the input text.
The GPT Conditioning and GPT Conditioning Chunk Length settings would be especially useful for users working with longer texts or experimenting with different voice generation behaviors.
It would be great if these options could be exposed as advanced settings in the voice generation interface, perhaps under an "Advanced Settings" section, while keeping the default values unchanged for users who don't need them.
Thank you again for the project, and I hope these controls can be considered for a future release.
Hello,
First of all, thank you for developing Voicebox.sh. I really appreciate the work that has gone into the project and the overall voice generation experience.
For a future update, I would like to suggest adding more advanced controls to the voice generation stage. It would be very useful if users could adjust the following parameters directly from the interface:
Speed
Temperature
Top P
Top K
Repetition Penalty
GPT Conditioning
GPT Conditioning Chunk Length
Having access to these parameters would give users more control over the generated speech and allow them to fine-tune the output for different use cases.
For example, parameters such as Temperature, Top P, Top K, and Repetition Penalty could help users control the generation behavior and improve pronunciation or stability in certain cases. Speed would also be useful for adjusting the speaking rate without having to modify the input text.
The GPT Conditioning and GPT Conditioning Chunk Length settings would be especially useful for users working with longer texts or experimenting with different voice generation behaviors.
It would be great if these options could be exposed as advanced settings in the voice generation interface, perhaps under an "Advanced Settings" section, while keeping the default values unchanged for users who don't need them.
Thank you again for the project, and I hope these controls can be considered for a future release.