Keras-oke

Exploring a pipeline for generating karaoke videos via Automatic Speech Recongnition inference

---

Dependencies:

• Spleeter

pip install spleeter 

for aeneas pipeline:

• Aeneas

sudo apt-get install libespeak-dev
pip install numpy
pip install aeneas

for whisper pipeline(deprecated): OpenAI whisper

pip install git+https://github.com/openai/whisper.git 

for correct whisper dependency versions, make sure you have CUDA >= 11.1 and run

pip install torchvision==0.11.0+cu113 torchaudio==0.10.0 -f https://download.pytorch.org/whl/torch_stable.html