APINET: A Methodology for Deploying Deep and Machine Learning Models as Services

The use of machine and deep learning models by external users is a complex task due to the large number of elements (hyperparameters, architectures, data types, etc.) that have to be known and applied. This work proposes a methodology for simple and efficient creation of APIs where users can access automatic learning models without knowing their implementation details. Tests and results with different neural network models and a learning environment show that the approach proposed eases the use of models created externally through a web service without compromising the output data quality.