# Audio Embedding with Neural Network Fingerprint

*Author: [Jael Gu](https://github.com/jaelgu)*

<br />

## Description

The audio embedding operator converts an input audio into a dense vector which can be used to represent the audio clip's semantics.
Each vector represents for an audio clip with a fixed length of around 1s.
This operator generates audio embeddings with fingerprinting method introduced by [Neural Audio Fingerprint](https://arxiv.org/abs/2010.11910).
The model is implemented in Pytorch.
We've also trained the nnfp model with [FMA dataset](https://github.com/mdeff/fma) (& some noise audio) and shared weights in this operator.
The nnfp operator is suitable for audio fingerprinting.

<br />

## Code Example

Generate embeddings for the audio "test.wav".

*Write a same pipeline with explicit inputs/outputs name specifications:*

- **option 1 (towhee>=0.9.0):**
```python
from towhee.dc2 import pipe, ops, DataCollection

p = (
    pipe.input('path')
        .map('path', 'frame', ops.audio_decode.ffmpeg())
        .map('frame', 'vecs', ops.audio_embedding.nnfp(device='cpu'))
        .output('path', 'vecs')
)

DataCollection(p('test.wav')).show()
```
<img src="./result.png" width="800px"/>

- **option 2:**
```python
import towhee

(
    towhee.glob['path']('test.wav')
          .audio_decode.ffmpeg['path', 'frames']()
          .runas_op['frames', 'frames'](func=lambda x:[y[0] for y in x])
          .audio_embedding.nnfp['frames', 'vecs']()
          .select['path', 'vecs']()
          .show()
)
```
<br />

## Factory Constructor

Create the operator via the following factory method

***audio_embedding.nnfp(model_name='nnfp_default', model_path=None, framework='pytorch')***

**Parameters:**

*model_name: str*

Model name to create nnfp model with different parameters.

*model_path: str*

The path to model. If None, it will load default model weights.

*framework: str*

The framework of model implementation.
Default value is "pytorch" since the model is implemented in Pytorch.

<br />

## Interface

An audio embedding operator generates vectors in numpy.ndarray given towhee audio frames.

***\_\_call\_\_(data)***

**Parameters:**

*data: List[towhee.types.audio_frame.AudioFrame]*

Input audio data is a list of towhee audio frames.
The audio input should be at least 1s.

**Returns**:

*numpy.ndarray*

Audio embeddings in shape (num_clips, 128).
Each embedding stands for features of an audio clip with length of 1s.

<br />

***save_model(format='pytorch', path='default')***

**Parameters:**

*format: str*

Format used to save model, defaults to 'pytorch'.
Accepted formats: 'pytorch', 'torchscript, 'onnx', 'tensorrt' (in progress)

*path: str*

Path to save model, defaults to 'default'.
The default path is under 'saved' in the same directory of operator cache.

```python
from towhee import ops

op = ops.audio_embedding.nnfp(device='cpu').get_op()
op.save_model('onnx', 'test.onnx')
```
    PosixPath('/Home/.towhee/operators/audio-embedding/nnfp/main/test.onnx')

## Fine-tune
To fine-tune this operator, please refer to [this example guide](https://github.com/towhee-io/examples/blob/main/fine_tune/7_fine_tune_audio_embedding_operator.ipynb).