Technique that Effectively Preserves Audio Identity in Bidirectional Style Conversion Using GAN

This is the code implementation for project submitted as the partial requirement for the 'Machine Learning Application' course in Korea University.

Instruction

python mkdir.py

python compile.py --istrain --istest

python gen_npy.py --iscut
python gen_npy.py --isnpy

python model.py

python target_inference.py

python compile.py --mkresult

python gen_spectrograms.py

python distance.py

Figure 4. in Paper

Table 1. in Paper

Name		Name	Last commit message	Last commit date
Latest commit History 2 Commits
.gitignore		.gitignore
README.md		README.md
compile.py		compile.py
config.py		config.py
distance.py		distance.py
figure1.png		figure1.png
gen_audio.py		gen_audio.py
gen_npy.py		gen_npy.py
gen_spectromgram.py		gen_spectromgram.py
mkdir.py		mkdir.py
model.py		model.py
target_inference.py		target_inference.py