Enhanced Progress Report

[MaxTheSpy]

Describe the feature
Would very much appreciate an enhanced progress report instead of just seconds and percentage.

Either a live Terminal view on the WebUI or to pull the location data 507/8914. GPU load would be a nice to have thing, as well as a live update percentage in the tab title that auto updates similar to Automatic1111 which also uses Gradio.

Screenshot schema or code
Provide a maximum of elements to help us to understand and create this new feature

[ROBERT-MCDOWELL]

terminal on gradio is planned since long time. despite of some integration I was not really happy of the result.
show the GPU load on gradio is ok for nvidia only. not for intel or AMD unless a subprocess calling rocm-smi which involves a lot of complication like install OS applications for Linux, Mac Windows and find a universal way to detect the GPU(s). this is something I worked on it months ago and too risky to satisfy every user with bug free... other priority are on rails for now like the quality of the TTS engine, new models, optimization of the conversion etc....

[MaxTheSpy]

terminal on gradio is planeed since long time. despite of some integration I was not really happy of the result. show the GPU load on gradio is ok for nvidia only. not for intel or AMD unless a subprocess calling rocm-smi which involves a lot of complication like install OS applications for Linux, Mac Windows and find a universal way to detect the GPU(s). this is something I worked on it months ago and too risky to satisfy every user with bug free... other priority are on rails for now like the quality of the TTS engine, new models, optimization of the conversion etc....

Thank you for being so on top of this! I can understand that. I think the terminal integration is the most important part, just to keep an eye on things.

I did look at your roadmap after making that post and I'm excited about the F5-TTS engine, Ive seen a few demos recently that sound really great!

Could you provide some information on the the current voice emulators? on how much data are they trained on? would it be better to use a voice cloning .wav thats like 20-30 minutes or longer?

[ROBERT-MCDOWELL]

Could you provide some information on the the current voice emulators? on how much data are they trained on? would it be better to > use a voice cloning .wav thats like 20-30 minutes or longer?

are you talking about coqui-tts or F5-TTS? if it's coqui-tts, a voicefile from 6 to 30 sec is enough.