MLWeb is an abstracted machine learning development platform, that supports both professional software engineers and domain-scientists across the data processing, model selection, and model training stages of a typical machine learning workflow. The platform aims to provide a simple interface for machine learning development that requires minimal user input, with human-friendly explanations to classification, and rapid testing and deployment of models to production.
The MLWeb APIs are 3 Docker containers acting as the backend of the platform, split into the tasks of model training, model prediction, and model explanation.
This repository contains the source code for the APIs, but the accompanying front-end source code can be found at craigmassie/mlweb-front. Both repositories were designed for a proof-of-concept evaluation as part of a MSci dissertation, and as such, may be a bit ugly in parts. Bear with me.