Fast piano transcription on AWS -Part 2

Previously, I tested a few architectures for music transcription. I concluded that separating the preprocessing, inferencing and postprocessing steps for parallel processing is crucial to improving model speed — refer to the architecture below: In t...