Executing Keras applications with LEMONADE on top of Kubernetes in a trustworthy infrastructure”, is a platform developed by the Federal University of Minas Gerais to create, test and run machine learning models, to create neural network models using Keras framework. Model building can be scheduled to run in a cloud infrastructure using ATMOSPHERE LEMONADE. Resource requirements, such as GPGPU support, can be defined in LEMONADE and are allocated by Kubernetes, allowing scale LEMONADE's modules to be scaled when demand increases. Finally, LEMONADE provides a set of tools to help data scientists to test and evaluate model's results regarding properties such as privacy, stability, fairness and transparency.
LEMONADE has been exposed as a service in the EOSC marketplace portal, which will increase its visibility and exploitation opportunities.
LEMONADE TPDS library will continue to be developed by Federal University of Minas Gerais and it is already being used in other projects both for the sake of education in Data Science and also as a platform for developing tools and applications
Federal University of Minas Gerais (Brazil), EOSC Portal (Europe), Keras applications (Worldwide)
Chief Technology Officer (CTO)
Specific needs / Value Proposition:
It consists of distributed federated services for big data. Users will be able to upload data sets using a service provided by Lemonade. Data are kept in a redundant file system, aimed to provide high-availability and high throughput.
Specific benefits / Value Proposition:
Handles heterogeneous data and provide high portability & interoperability. Data storage requirements will depend on use cases and installation. Users may process terabytes of data and their volume will directly impact the storage and processing costs. Lemonade can be scaled to support hundreds of users by increasing cluster capacity. A large number of users can be supported in a modest cluster of commodity computers and a volume of data often found in most of organizations.