Spaces:
Running
Running
Upload README.md with huggingface_hub
Browse files
README.md
CHANGED
|
@@ -1,15 +1,61 @@
|
|
| 1 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2 |
sdk: gradio
|
| 3 |
-
|
| 4 |
-
emoji: 👀
|
| 5 |
-
colorFrom: yellow
|
| 6 |
-
colorTo: gray
|
| 7 |
-
sdk_version: 6.14.0
|
| 8 |
-
python_version: '3.13'
|
| 9 |
-
app_file: app.py
|
| 10 |
pinned: false
|
| 11 |
-
tags:
|
| 12 |
-
- ml-intern
|
| 13 |
---
|
| 14 |
|
| 15 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
+
title: VoiceForge Studio
|
| 3 |
+
emoji: 🎙️
|
| 4 |
+
colorFrom: indigo
|
| 5 |
+
colorTo: purple
|
| 6 |
sdk: gradio
|
| 7 |
+
app_port: 7860
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 8 |
pinned: false
|
|
|
|
|
|
|
| 9 |
---
|
| 10 |
|
| 11 |
+
# 🎙️ VoiceForge Studio
|
| 12 |
+
|
| 13 |
+
A professional, modern web application for Text-to-Speech (TTS) and Automatic Speech Recognition (ASR) powered by the latest open-source models.
|
| 14 |
+
|
| 15 |
+
## Features
|
| 16 |
+
|
| 17 |
+
### 🗣️ Text-to-Speech
|
| 18 |
+
- **Voice Cloning** — Clone any voice from a reference audio sample
|
| 19 |
+
- **Voice Design** — Create unique voices from text descriptions
|
| 20 |
+
- **Emotion Control** — Adjust emotional tone (happy, sad, angry, etc.)
|
| 21 |
+
- **Multi-speaker** — Generate conversations with multiple speakers
|
| 22 |
+
- **Multiple Models**: Qwen3-TTS, VibeVoice, Voxtral, IndexTTS-2, OmniVoice
|
| 23 |
+
|
| 24 |
+
### 🎙️ Speech-to-Text (ASR)
|
| 25 |
+
- **Word-Level Timestamps** — Precise timing for every single word
|
| 26 |
+
- **Line-Level Timestamps** — Chunk-level timing for readability
|
| 27 |
+
- **Speaker Diarization** — Identify who spoke when (pyannote 3.1)
|
| 28 |
+
- **Speaker Identification** — Automatic speaker counting and labeling
|
| 29 |
+
- **Multiple Models**: Whisper Turbo, Qwen3-ASR, VibeVoice-ASR
|
| 30 |
+
|
| 31 |
+
### 📦 Batch Processing
|
| 32 |
+
- Process multiple audio files at once
|
| 33 |
+
- Export results as JSON
|
| 34 |
+
- Automatic diarization across all files
|
| 35 |
+
|
| 36 |
+
## Models Supported
|
| 37 |
+
|
| 38 |
+
| Type | Model | Size | Key Features |
|
| 39 |
+
|------|-------|------|-------------|
|
| 40 |
+
| TTS | Qwen3-TTS | 1.7B | Voice Design + Clone, 12 languages |
|
| 41 |
+
| TTS | VibeVoice 1.5B | 1.5B | Long-form 90min, 4 speakers |
|
| 42 |
+
| TTS | VibeVoice Realtime | 0.5B | Streaming, low latency |
|
| 43 |
+
| TTS | Voxtral 4B | 4B | vLLM, OpenAI-compatible |
|
| 44 |
+
| TTS | IndexTTS-2 | — | Emotion control (8 dims) |
|
| 45 |
+
| TTS | OmniVoice | — | OpenAI-compatible API |
|
| 46 |
+
| ASR | Whisper Turbo | — | Word timestamps, 99 langs |
|
| 47 |
+
| ASR | Qwen3-ASR 1.7B | 1.7B | Forced alignment, line stamps |
|
| 48 |
+
| ASR | Qwen3-ASR 0.6B | 0.6B | Fast, lightweight |
|
| 49 |
+
| ASR | VibeVoice-ASR | — | Native diarization + timestamps |
|
| 50 |
+
| Diarization | pyannote 3.1 | — | Speaker diarization |
|
| 51 |
+
|
| 52 |
+
## Usage
|
| 53 |
+
|
| 54 |
+
1. Select a TTS or ASR model from the dropdown
|
| 55 |
+
2. For TTS: Enter text, optionally upload a reference voice for cloning
|
| 56 |
+
3. For ASR: Upload audio, choose timestamp granularity and diarization options
|
| 57 |
+
4. Click Generate / Transcribe and view the beautiful output
|
| 58 |
+
|
| 59 |
+
## License
|
| 60 |
+
|
| 61 |
+
Models follow their respective licenses (MIT, Apache-2.0, etc.)
|