# NovaSR

**Repository Path**: baxtax/NovaSR

## Basic Information

- **Project Name**: NovaSR
- **Description**: No description available
- **Primary Language**: Unknown
- **License**: Apache-2.0
- **Default Branch**: main
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 1
- **Created**: 2026-05-16
- **Last Updated**: 2026-05-16

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

## NovaSR: Pushing the Limits of Extreme Efficiency in Audio Super-Resolution

<p align="center">
  <a href="https://huggingface.co/YatharthS/NovaSR">
    <img src="https://img.shields.io/badge/%F0%9F%A4%97%20Hugging%20Face-Model-FFD21E" alt="Hugging Face Model">
  </a>
  &nbsp;
  <a href="https://huggingface.co/spaces/YatharthS/NovaSR">
    <img src="https://img.shields.io/badge/%F0%9F%A4%97%20Hugging%20Face-Space-blue" alt="Hugging Face Space">
  </a>
  &nbsp;
  <a href="https://www.kaggle.com/code/yatharthsharma888/novasr-training">
    <img src="https://img.shields.io/badge/Kaggle-Training%20Notebook-20BEFF?logo=kaggle&logoColor=white" alt="Kaggle Notebook">
  </a>
</p>


This is the repository for NovaSR, a tiny 50kb audio upsampling model that upscales muffled 16khz audio into clear and crisp 48khz audio at speeds over 3500x realtime. 


https://github.com/user-attachments/assets/c81f87eb-f6de-4bf9-85bd-dfc9a223a865


### Key benefits
* Speed: Can reach 3600x realtime speed on a single a100 gpu.
* Quality: On par with models 5,000x larger.
* Size: Just 52kb in size, several thousand times smaller then most. 

### Why is this even useful?
* Enhancing models: NovaSR can enhance TTS model quality considerably with nearly 0 computational cost.
* Real-time enhancement: NovaSR allows for on device enhancement of any low quality calls, audio, etc. while using nearly no memory.
* Restoring datasets: NovaSR can enhance audio quality of any audio dataset. 


### Comparisons

Comparisons were done on A100 gpu. Higher realtime means faster processing speeds.

| Model         | Speed (Real-Time) | Model Size |
| :------------ | :---------------- | :--------- |
| **NovaSR** | **3600x realtime** | **~52 KB** |
| FlowHigh      | 20x realtime        | ~450 MB     |
| FlashSR       | 14x realtime        | ~1000 MB     |
| AudioSR       | 0.6x realtime    | ~2000 MB     |

### Examples
Please check the [huggingface model](https://huggingface.co/YatharthS/NovaSR) for a few examples. 



### Usage
You can try it on [huggingface spaces](https://huggingface.co/spaces/YatharthS/NovaSR) or locally.


Simple 1 line installation:
```
pip install git+https://github.com/ysharma3501/NovaSR.git
```

Load model
```python
from NovaSR import FastSR

upsampler = FastSR() ## downloads from hf

## Use this instead for CPUs as it leads to 3-4x speedup.
# upsampler = FastSR(half=False)
```

Run model
```python
from IPython.display import Audio

## replace audio_path.wav with your wav/mp3 file
lowres_audio = upsampler.load_audio('audio_path.wav') 

## infer with model
highres_audio = upsampler.infer(lowres_audio).cpu()

display(Audio(highres_audio, rate=48000))
```
### Training

Please check out the kaggle notebook for training the model further on custom datasets: https://www.kaggle.com/code/yatharthsharma888/novasr-training
### Info

Q: How much data was this trained on?

A: Just 100 hours of data(mls_sidon along with vctk)

Q: How is it so small?

A: It uses less then 10 tiny conv1d layers along with snake activations based on bigvgan for maximum quality and size.

Q: Will benchmarks come?

A: Yes, I am still training it further and will benchmark it later.

## Final Notes

Repo stars and model likes would be appreciated if found helpful, thank you.


Email: yatharthsharma3501@gmail.com