Metadata-Version: 2.1
Name: rvc-python
Version: 0.0.1
Summary: Use RVC via console or python scripts
Project-URL: Homepage, https://github.com/daswer123/rvc-python
Project-URL: Bug Tracker, https://github.com/daswer123/rvc-python/issues
Author-email: daswer123 <daswerq123@gmail.com>
License-File: LICENSE
Classifier: License :: OSI Approved :: MIT License
Classifier: Operating System :: OS Independent
Classifier: Programming Language :: Python :: 3
Requires-Python: >=3.9
Requires-Dist: av
Requires-Dist: fairseq==0.12.2
Requires-Dist: faiss-cpu
Requires-Dist: ffmpeg-python
Requires-Dist: loguru
Requires-Dist: numpy<=1.25.3
Requires-Dist: omegaconf==2.0.6
Requires-Dist: praat-parselmouth>=0.4.2
Requires-Dist: pydub
Requires-Dist: pyworld
Requires-Dist: requests==2.31.0
Requires-Dist: soundfile
Requires-Dist: torch>=2.0.0
Requires-Dist: torchcrepe
Description-Content-Type: text/markdown

# rvc-python

Using RVC via console or python scripts

Feel free to make PRs or use the code for your own needs

## Changelog

You can keep track of all changes on the [release page](https://github.com/daswer123/https://github.com/daswer123/rvc-python/releases)

## Installation

Simple installation :

```bash
pip install rvc-python
```

This will install all the necessary dependencies, including a **CPU support only** version of PyTorch

I recommend that you install the **GPU version** to improve processing speed ( up to 3 times faster )

### Windows
```bash
python -m venv venv
venv\Scripts\activate
pip install rvc-python
pip install torch==2.1.1+cu118 torchaudio==2.1.1+cu118 --index-url https://download.pytorch.org/whl/cu118
```

### Linux
```bash
python -m venv venv
source venv\bin\activate
pip install rvc-python
pip install torch==2.1.1+cu118 torchaudio==2.1.1+cu118 --index-url https://download.pytorch.org/whl/cu118
```

## Usage

python -m rvc_python [-h] [-pi PITCH] [-i INPUT] [-ip INDEX] [-me METHOD] [-v VERSION] [-o OUTPUT]
                  [-mp MODEL] [-ir INDEX_RATE] [-d DEVICE]
                  [-fr FILTER_RADIUS]  [-rsr RESAMPLE_SR]
                  [-rmr RMS_MIX_RATE][-pr PROTECT]

Process and convert voices using specified models and methods.

options:
  -h, --help            show this help message and exit
  -pi PITCH, --pitch PITCH
                        Transpose integer, number of semitones. Raise by an octave: +12,
                        lower by an octave: -12.
  -i INPUT, --input INPUT
                        Path to input file (mandatory).
  -ip INDEX, --index INDEX
                        Path to index file (optional).
  -me METHOD, --method METHOD
                        Pitch extraction algorithm choices: ['harvest', 'crepe', 'rmvpe', 'pm'].
  -v VERSION, --version VERSION
                        Version of the software or model ('v1' or 'v2').
  -o OUTPUT, --output OUTPUT
                        Output path for results (default is "out.wav").
  -mp MODEL, --model MODEL
                        Path to model file (mandatory).
  -ir INDEX_RATE, --index_rate INDEX_RATE
                        Search feature ratio.
  -d DEVICE, --device DEVICE
                       Device to use for processing (e.g., 'cpu' or 'cuda:0').
  -fr FILTER_RADIUS, --filter_radius FILTER_RADIUS
                       Apply median filtering to pitch results. Represents filter radius.
                       A larger value can reduce breathiness.
  -rsr RESAMPLE_SR ,--resample_sr RESAMPLE_SR
                      Resample output audio in post-processing. Set to "0" for no resampling.

  	-rmr RMS_MIX_RATE ,--rms_mix_rate RMS_MIX_RATE
                      Mix rate between input volume envelope and output volume envelope. Closer to "1" uses more from output.

  	-pr PROTECT ,--protect PROTECT
                      Protects voiceless consonants and breath sounds from artifacts such as tearing in electronic music. Decrease value for increased protection but may affect indexing accuracy.

