README.md

metadata

title: PersonaPlex
emoji: 🎭
colorFrom: purple
colorTo: blue
sdk: gradio
sdk_version: 6.3.0
app_file: app.py
pinned: false
hardware: zerogpu
python_version: '3.10'

PersonaPlex 7B Demo

Interactive demo for nvidia/personaplex-7b-v1 - a multimodal speech-to-speech model capable of real-time persona-driven conversation.

Features

• Voice Input: Record or upload audio
• Persona Selection: Choose from different conversation personas
• Voice Cloning: Select different voice styles for output
• Real-time Generation: Streaming speech generation

Usage

1. Record or upload an audio clip
2. Select a persona (e.g., "helpful assistant", "casual friend")
3. Choose an output voice
4. Click Generate to hear the response

Model Info

PersonaPlex is based on the Moshi architecture and supports:

• Audio-to-audio generation
• Persona conditioning
• Multiple voice embeddings
• Streaming inference

Requirements

This Space requires access to the gated model. Make sure you have accepted the license at nvidia/personaplex-7b-v1.