update readme

README.md

@@ -1,58 +1,85 @@
 ---
 license: mit
 language:
-- pt
+- multilingual
 tags:
-- chatterbox
 - text-to-speech
 - tts
 - multilingual
-- single-language-tts
 - voice-cloning
+- speech-synthesis
+- chatterbox
 - chatterbox-v3
 pipeline_tag: text-to-speech
 base_model: ResembleAI/chatterbox
-base_model_relation: finetune
 ---
 
-<!-- chatterbox-space-link -->
-> 🎙️ **Live demo:** Try this model in the [`ResembleAI/Chatterbox-Multilingual-TTS-pt-br`](https://huggingface.co/spaces/ResembleAI/Chatterbox-Multilingual-TTS-pt-br) Space.
-<!-- chatterbox-space-link -->
+# SonicVox-Multilingual
+
+🎙️ **SonicVox-Multilingual** is a multilingual Text-to-Speech (TTS) model
+built on top of **Chatterbox Multilingual** by Resemble AI.
+
+It is designed to generate natural, expressive speech across multiple
+languages with support for **zero-shot voice cloning** using a reference
+audio sample.
+
+## ✨ Features
 
-# Chatterbox Multilingual: Brazilian Portuguese
+- 🌍 Multilingual Text-to-Speech
+- 🎙️ Natural and expressive speech generation
+- 🗣️ Zero-shot voice cloning
+- 🔊 High-quality speech synthesis
+- 🎚️ Voice conditioning using reference audio
+- ⚡ GPU-accelerated inference
+- 🧩 Based on Chatterbox Multilingual V3
 
-Chatterbox Multilingual: Brazilian Portuguese is a dedicated single-language finetune in the **Chatterbox Multilingual V3 Single Language Pack**. It is optimized for Portuguese as spoken in Brazil, with language- and region-specific behavior for expressive text-to-speech and voice cloning.
+## 🌎 Supported Languages
 
-Use this model when you want tighter Brazilian Portuguese quality control than the broad multilingual checkpoint. For a single model that covers all supported languages, use [`ResembleAI/chatterbox`](https://huggingface.co/ResembleAI/chatterbox).
+SonicVox-Multilingual is based on the multilingual Chatterbox model.
 
-## Demo
+Supported languages depend on the underlying checkpoint and include
+multiple languages such as:
 
-Try the hosted demo Space: [`ResembleAI/Chatterbox-Multilingual-TTS-pt-br`](https://huggingface.co/spaces/ResembleAI/Chatterbox-Multilingual-TTS-pt-br).
+- English
+- Spanish
+- French
+- German
+- Italian
+- Portuguese
+- Hindi
+- Telugu
+- Tamil
+- Kannada
+- Malayalam
+- Chinese
+- Japanese
+- Korean
+- Arabic
+- Russian
+- Turkish
+- Polish
 
-## Files
+> Always verify the supported language IDs with the specific Chatterbox
+> checkpoint being used.
 
-- `t3_pt_br.safetensors`: T3 state dict in safetensors format.
-- `s3gen_v3.pt` / `s3gen_v3.safetensors`: V3 S3Gen speech decoder checkpoint.
-- `grapheme_mtl_merged_expanded_v1.json`: multilingual tokenizer config.
+## 🎤 Voice Cloning
 
-## Language
+SonicVox-Multilingual supports voice cloning using a reference audio file.
 
-- Locale: `pt-BR`
-- Chatterbox language ID: `pt`
+Example:
 
-## Checkpoint Metadata
+```python
+import torchaudio as ta
+from chatterbox.mtl_tts import ChatterboxMultilingualTTS
 
-- Source step: `137400`
-- Source checkpoint: `t3_137400.pth.tar`
-- Tensor count: `292`
-- Dtype: `float32`
-- Text embedding shape: `(2454, 1024)`
-- Speech embedding shape: `(8194, 1024)`
-- Size: `2143990296` bytes
-- SHA256: `074aaf65255eb9cb960288f7cc72e09d3b5008f6e0b14868c0d4e5b0bd7cbb6c`
+model = ChatterboxMultilingualTTS.from_pretrained(device="cuda")
 
-## Loader Notes
+text = "Welcome to SonicVox Multilingual."
 
-This repository contains Chatterbox Multilingual V3 single-language assets used by the linked demo Space. The T3 checkpoint is loaded with multilingual vocabulary shape `2454` and S3 speech vocabulary shape `8194`.
+wav = model.generate(
+    text,
+    language_id="en",
+    audio_prompt_path="reference.wav"
+)
 
-The demo combines these model-specific assets with the shared Chatterbox inference code and companion assets needed for end-to-end speech generation.
+ta.save("output.wav", wav, model.sr)