Skip to content

Commit

Permalink
Update instructions to install from PyPi
Browse files Browse the repository at this point in the history
  • Loading branch information
juanmc2005 committed Dec 15, 2021
1 parent 0b500c9 commit 5894ad6
Showing 1 changed file with 11 additions and 5 deletions.
16 changes: 11 additions & 5 deletions README.md
Original file line number Diff line number Diff line change
@@ -1,4 +1,6 @@
# Streaming speaker diarization
# Diart - Streaming speaker diarization in real time

[![PyPI version](https://badge.fury.io/py/diart.svg)](https://badge.fury.io/py/diart)

*[Overlap-aware low-latency online speaker diarization based on end-to-end local segmentation](/paper.pdf)*
by [Juan Manuel Coria](https://juanmc2005.github.io/), [Hervé Bredin](https://herve.niderb.fr), [Sahar Ghannay](https://saharghannay.github.io/) and [Sophie Rosset](https://perso.limsi.fr/rosset/).
Expand Down Expand Up @@ -27,10 +29,14 @@ conda activate diarization

2) Install the latest PyTorch version following the [official instructions](https://pytorch.org/get-started/locally/#start-locally)

3) Install:
3) Install pyannote.audio 2.0 (currently in development)
```shell
pip install git+https://github.com/pyannote/pyannote-audio.git@develop#egg=pyannote-audio
```

4) Install Diart:
```shell
cd StreamingSpeakerDiarization
pip install -e .
pip install diart
```

## Usage
Expand All @@ -51,7 +57,7 @@ python -m diart.demo microphone

This will launch a real-time visualization of the diarization outputs as they are produced by the system:

![Example of a state of the real-time output plot](/visualization.png)
![Example of a state of the real time output plot](/visualization.png)

By default, the script uses step = latency = 500ms, and it sets reasonable values for all hyper-parameters.
See `python -m diart.demo -h` for more information.
Expand Down

0 comments on commit 5894ad6

Please sign in to comment.