Transformer Groove Infilling implementation, part of my Master Thesis: Completing Audio Drum Loops with Transformer Neural Networks, under the supervision of Sergi Jordà and Behzad Haki at the Music Technology Group at the Universitat Pompeu Fabra (Barcelona).

Object-oriented micro-service implemented in Python based on Deep Learning model BERT. --- This is a micro-service exposed via the FastAPI that allows to infill words/sentences into a given context based on a paper by Stanford University. The corresponding article got featured on medium.com/analytics-vidhya.

Multivariate time series infilling using the Normal copula as the dependence model. Data plausibility check by comparing known data against infilled values. Various plotting routines to visualize results.