1 code implementation • 9 Nov 2023 • Johannes Hagemann, Samuel Weinbach, Konstantin Dobler, Maximilian Schall, Gerard de Melo
In this work, we conduct a comprehensive ablation study of possible training configurations for large language models.
2 code implementations • 23 May 2023 • Konstantin Dobler, Gerard de Melo
However, if we want to use a new tokenizer specialized for the target language, we cannot transfer the source model's embedding matrix.
1 code implementation • 23 Feb 2022 • Konstantin Dobler, Florian Hübscher, Jan Westphal, Alejandro Sierra-Múnera, Gerard de Melo, Ralf Krestel
Our approach is based on the StyleGAN neural network architecture, but incorporates a custom multi-conditional control mechanism that provides fine-granular control over characteristics of the generated paintings, e. g., with regard to the perceived emotion evoked in a spectator.