Copy the voice of anyone using machine learning.
- Install following dependencies
sudo apt-get install -y git csh make g++ unzip tree
sudo apt-get install -y libcurl4-openssl-dev
sudo apt-get install -y python-virtualenv
sudo apt-get install -y python-dev
- Setup Virtual environment
cd ~/mimicry.ai
virtualenv venv
source venv/bin/activate
- Install Python dependencies
pip install --upgrade pip
pip install numpy scipy matplotlib ipython jupyter pandas sympy nose
pip install theano lxml
export CFLAGS=-I/home/vagrant/mimicry.ai/venv/lib/python2.7/site-packages/numpy/core/include
pip install bandmat
- Install dependencies
cd ~/mimicry.ai/src/merlin/tools
/bin/bash compile_tools.sh
- Run test (Optional. Not needed)
cd ~/mimicry.ai/src/merlin/test
/bin/bash test_training.sh
- Run demo
cd ~/mimicry.ai
source venv/bin/activate
cd ~/mimicry.ai/src/merlin/egs/slt_arctic/s1
./run_demo.sh
-
Install dependencies
-
Find voice of top actors, politicians
- Rajini
- Kamalkasan
- Modi
- Jayalalitha
- P Sidambaram
- Vadivelu
- Karunanithi
-
Train the machine with sample data
-
Setup Sample UI
-
Provide text for conversion
https://pagalworld.me/ http://mp3lier.info/
Tools:
youtube-dl --extract-audio --audio-format mp3 https://www.youtube.com/watch?v=4CxkK1gam8o
-
Install dependencies
cd ~/mimicry.ai source venv/bin/activate
cd server pip install flask
Run Python server
python index.py
Request:
curl -XPOST http://api.mimicry.ai/mimic -F 'speaker=modi' -F 'content=I Love India'