VibeVoice issues

fix: issues in error handling

1

--- ## 🔄 Pull Request (Concise) ```markdown # fix: WebSocket error handling causes crashes and log pollution ## 📋 Summary Fixes critical error handling in WebSocket streaming that causes crashes...

ikeshav26

Could you provide a Chinese speaker?

1

Dear dev: The effect of this project is excellent. Could you provide a Chinese speaker? BRs

nianliunian

Large text problem

2

This is giving me 500 error when I gave it a large text like an article(3,584 tokens) . this should work because this is not that much token .

Priyabhunia

Question: Realtime model inference and training implementation details

10

Hi. Im trying to understand whats happens inside inference loop, and there one thing which is not clear for me. For diffusion model you are using CFG by two inputs...

alxmamaev

## Implementation of Quantization Support Following the maintainers' guidance to **contribute code rather than documentation**, this PR implements the quantization feature discussed previously. **Problem:** VibeVoice requires 20GB VRAM, blocking 90%...

maitrisavaliya

Feature Request: Addition of Unique/Stylized Voice Tones (e.g., Cute Bot, Anime)

3

Hello VibeVoice Team, I hope this is the appropriate channel for a feature request. I am currently developing a personal project, specifically a small desk robot, and I believe VibeVoice...

dw4n

Input Audio files words are being in output instead of my text prompt

I have an input audio file which is the cloning voice. I connected it into the Multi Speaker K Sampler and then i take the audio and save it from...

koloboyy

fix: compatibilidad GTX 1060 (sm_61) y modo host

Resumen:\n- Añade DEBIAN_FRONTEND=noninteractive al Dockerfile generado.\n- Manejo defensivo de RuntimeError en cierre de WebSocket (demo/web/app.py).\n- El generador systemd respeta VIBE_API_MODE=host para evitar arrancar la API en Docker.\n- Documentación: LEEME y...

jhoavera