ebellob/qwav3_4B

Model Card for Qwav3-4B

Qwav3-4B is an LLM-based text-to-speech (TTS) model that uses the WavTokenizer neural codec to synthesise speech. It has been pretrained with various datasets on Spanish, Catalan and English speech synthesis and zero-shot voice cloning.

The training recipe involves a VALL-E approach of giving 400 audio tokens as context and then having the model perform zero-shot voice cloning. This has not been completely accomplished, as the amount of training data is insufficient. However, the model is capable of grasping concepts such as the gender of the voice, producing male voices when given male context and viceversa.

Training and inference recipes will be published soon.

Model Details

Model Description

This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.

• Developed by: Erik Beltran Lobato (ebellob)
• Model type: Text-to-Speech LLM
• Language(s) (NLP): English, Spanish and Catalan
• License: [More Information Needed]
• Finetuned from model [optional]: Qwen3-4B-Instruct-2507 https://huggingface.co/Qwen/Qwen3-4B-Instruct-2507

Model Sources [optional]

• Repository: TODO

Uses

Direct Use

[More Information Needed]

Downstream Use [optional]

[More Information Needed]

Out-of-Scope Use

[More Information Needed]

Bias, Risks, and Limitations

[More Information Needed]

Recommendations

Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.

How to Get Started with the Model

Use the code below to get started with the model.

[More Information Needed]

Training Details

Training Data

TODO [More Information Needed]

Training Procedure

TODO

Preprocessing [optional]

[More Information Needed] TODO

Training Hyperparameters

• Training regime: bf16 with lr = 1e-4

Speeds, Sizes, Times [optional]

TODO

[More Information Needed]

Evaluation

TODO

Testing Data, Factors & Metrics

Testing Data

[More Information Needed]

Factors

[More Information Needed]

Metrics

[More Information Needed]

Results

[More Information Needed]

Summary

Model Examination [optional]

[More Information Needed]

Environmental Impact

Carbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).

• Hardware Type: Nvidia A40 GPU
• Hours used: 72h

Technical Specifications [optional]

Model Architecture and Objective

[More Information Needed]

Compute Infrastructure

[More Information Needed]

Hardware

[More Information Needed]

Software

[More Information Needed]

Citation [optional]

WavTokenizer: https://arxiv.org/abs/2408.16532 Qwen3: https://arxiv.org/abs/2505.09388 Datasets: TODO

BibTeX:

[More Information Needed]

APA:

[More Information Needed]

Glossary [optional]

[More Information Needed]

More Information [optional]

[More Information Needed]

Model Card Authors [optional]

[More Information Needed]

Model Card Contact

[More Information Needed]