# Model Card for Diva Llama 3

<!-- Provide a quick summary of what the model is/does. [Optional] -->
This is an end-to-end Voice Assistant Model which can handle speech and text as inputs. It is trained using distillation loss. More details will be in a paper [COMING SOON]!

See the model in action compared to SALMONN and Qwen-Audio at [diva-audio.github.io](https://diva-audio.github.io).
## Citation
No Publication As of Yet, But If You Use Please Cite the Below
**BibTeX:**

```
	@misc{held2024diva,
	  author="Held, Will and Zhang, Yanzhe and Ryan, Michael and Shi, Weiyan and Li, Ella and Yang, Diyi",
	  title="Distilling an End-to-End Voice Assistant from Speech Recognition Data",
	  year="2024",
	  publisher="HuggingFace",
	}
    
```

##  Table of Contents

- [Model Card for DiVA Llama 3](#model-card-for-DiVA-Llama-3)
- [Citation](#citation)
- [Table of Contents](#table-of-contents)
- [Training Details](#training-details)
  - [Training Data](#training-data)
  - [Training Procedure](#training-procedure)
- [Environmental Impact](#environmental-impact)
- [Technical Specifications [optional]](#technical-specifications-optional)
  - [Model Architecture and Objective](#model-architecture-and-objective)
  - [Compute Infrastructure](#compute-infrastructure)
    - [Hardware](#hardware)
    - [Software](#software)
- [Model Card Contact](#model-card-contact)

## Training Details

### Training Data

<!-- This should link to a Data Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->

This model was trained on the [CommonVoice](https://huggingface.co/datasets/mozilla-foundation/common_voice_16_1) corpus.


### Training Procedure

This model was trained for 7k gradient steps with a batch size of 512 Recordings and a linearly decaying learning rate from 5e-5 to zero, with a linear warmup of 70 steps.

### Environmental Impact

- **Hardware Type:** V4-32 TPU
- **Hours used:** 8 Hours
- **Cloud Provider:** Google Cloud.
- **Compute Region:** US Central C


### Hardware

This model was trained on at V4 TPU on Google Cloud.

### Software

This model was trained with [Levanter](https://github.com/stanford-crfm/levanter) 


## Model Card Authors [optional]

<!-- This section provides another layer of transparency and accountability. Whose views is this model card representing? How many voices were included in its construction? Etc. -->

Will Held

## Model Card Contact

held@stanford.edu