---
base_model:
- facebook/mms-lid-256
language:
- fr
license: cc-by-nc-4.0
metrics:
- accuracy
pipeline_tag: audio-classification
tags:
- model_hub_mixin
- pytorch_model_hub_mixin
- speaker_dialect_classification
library_name: transformers
---

# MMS-LID-256 for French Dialect Classification

# Model Description
This model includes the implementation of French dialect classification described in <a href="https://arxiv.org/abs/2508.01691"><strong>**Voxlect: A Speech Foundation Model Benchmark for Modeling Dialect and Regional Languages Around the Globe**</strong></a>

Github repository: https://github.com/tiantiaf0627/voxlect

The included French dialects are with speakers from: 
```
[
  "Africa",
  "France",
  "Canada",
  "Swiss/Belgium/German"
]
```

# How to use this model

## Download repo
```bash
git clone git@github.com:tiantiaf0627/voxlect
```
## Install the package
```bash
conda create -n voxlect python=3.8
cd voxlect
pip install -e .
```

## Load the model
```python
# Load libraries
import torch
import torch.nn.functional as F
from src.model.dialect.mms_dialect import MMSWrapper

# Find device
device = torch.device("cuda") if torch.cuda.is_available() else "cpu"

# Load model from Huggingface
model = MMSWrapper.from_pretrained("tiantiaf/voxlect-french-dialect-mms-lid-256").to(device)
model.eval()
```

## Prediction
```python
# Label List
dialect_list = [
  "Africa",
  "France",
  "Canada",
  "Swiss/Belgium/German"
]
    
# Load data, here just zeros as an example
# Our training data filters output audio shorter than 3 seconds (unreliable predictions) and longer than 15 seconds (computation limitation)
# So you need to prepare your audio to a maximum of 15 seconds, 16kHz, and mono channel
max_audio_length = 15 * 16000
data = torch.zeros([1, 16000]).float().to(device)[:, :max_audio_length]
logits, embeddings = model(data, return_feature=True)
    
# Probability and output
dialect_prob = F.softmax(logits, dim=1)
print(dialect_list[torch.argmax(dialect_prob).detach().cpu().item()])
```

Responsible Use: Users should respect the privacy and consent of the data subjects, and adhere to the relevant laws and regulations in their jurisdictions when using Voxlect.

## If you have any questions, please contact: Tiantian Feng (tiantiaf@usc.edu)

❌ **Out-of-Scope Use**
- Clinical or diagnostic applications
- Surveillance
- Privacy-invasive applications
- No commercial use

#### If you like our work or use the models in your work, kindly cite the following. We appreciate your recognition!
```
@article{feng2025voxlect,
  title={Voxlect: A Speech Foundation Model Benchmark for Modeling Dialects and Regional Languages Around the Globe},
  author={Feng, Tiantian and Huang, Kevin and Xu, Anfeng and Shi, Xuan and Lertpetchpun, Thanathai and Lee, Jihwan and Lee, Yoonjeong and Byrd, Dani and Narayanan, Shrikanth},
  journal={arXiv preprint arXiv:2508.01691},
  year={2025}
}
```