ResembleAI/Chatterbox-Multilingual-pt-br

1---2license: mit3language:4- pt5tags:6- chatterbox7- text-to-speech8- tts9- multilingual10- single-language-tts11- voice-cloning12- chatterbox-v313pipeline_tag: text-to-speech14base_model: ResembleAI/chatterbox15base_model_relation: finetune16---17 18<!-- chatterbox-space-link -->19> 🎙️ **Live demo:** Try this model in the [`ResembleAI/Chatterbox-Multilingual-TTS-pt-br`](https://huggingface.co/spaces/ResembleAI/Chatterbox-Multilingual-TTS-pt-br) Space.20<!-- chatterbox-space-link -->21 22# Chatterbox Multilingual: Brazilian Portuguese23 24Chatterbox 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.25 26Use 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).27 28## Demo29 30Try the hosted demo Space: [`ResembleAI/Chatterbox-Multilingual-TTS-pt-br`](https://huggingface.co/spaces/ResembleAI/Chatterbox-Multilingual-TTS-pt-br).31 32## Files33 34- `t3_pt_br.safetensors`: T3 state dict in safetensors format.35- `s3gen_v3.pt` / `s3gen_v3.safetensors`: V3 S3Gen speech decoder checkpoint.36- `grapheme_mtl_merged_expanded_v1.json`: multilingual tokenizer config.37 38## Language39 40- Locale: `pt-BR`41- Chatterbox language ID: `pt`42 43## Checkpoint Metadata44 45- Source step: `137400`46- Source checkpoint: `t3_137400.pth.tar`47- Tensor count: `292`48- Dtype: `float32`49- Text embedding shape: `(2454, 1024)`50- Speech embedding shape: `(8194, 1024)`51- Size: `2143990296` bytes52- SHA256: `074aaf65255eb9cb960288f7cc72e09d3b5008f6e0b14868c0d4e5b0bd7cbb6c`53 54## Loader Notes55 56This 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`.57 58The demo combines these model-specific assets with the shared Chatterbox inference code and companion assets needed for end-to-end speech generation.59