Training a Speech-to-Text Model for Dutch on the Corpus Gesproken Nederlands
This talk will be about my paper with the same title. I will go through the preprocessing methods we used for the corpus. Afterwards I will dive into the creation of our first set of models and their performance. Lastly, I will talk about how transfer learning allowed us to leverage the power of a well performing English model, for our Dutch models.
Willem Röpke is a first-year master’s student at the Vrije Universiteit Brussel, with a special interest in Artificial Intelligence. When he is not studying or spending time with friends and family, he likes to dedicate his time to a non-profit that he helps lead.