no code implementations • 15 Aug 2022 • Aaron Valero Puche, Sukhan Lee
In spite of the fact that AlphaGo versions are known to be computationally heavy, we manage to train the proposed framework with a single thread and GPU, while obtaining a solution that outperforms the state-of-the-art results in space utilization.
1 code implementation • IEEE MLSP 2021 • Aaron Valero Puche, Sukhan Lee
We show that training a conditional autoencoder based on accuracy in timbre classification together with adversarial regularization of pitch content allows timbre distribution in latent space to be more effective and stable for timbre interpolation and pitch conditioning.