There is a variety of available open source technologies for planning and implementing experiments in Opentech AI R&D. This page collects links to some of the platforms and technologies used in R&D of AI systems.
- TensorFlow is an open source platform for expressing and experimenting with machine learning ideas and systems.
- Keras is a high-level neural networks API capable of running on TensorFlow, CNTK or Theano.
- PyTorch is a open source deep learning framework for tensors and neural networks processing in Python with GPU acceleration support.
- Anaconda is a widely used open source Python distribution for data science. It includes over 1000 data science packages and a set of IDEs (Jupyter, Spyder, RStudio) and visualization tools.
- Microsoft Cognitive Toolkit (CNTK) is an open source deep-learning toolkit .
- Theano is Python library that can be used for defining, optimizing and evaluating mathematical models in machine/deep learning. Recent announcement (10/1/2017) states that this historic package is being discontinued for regular updates.
- Machine Learning as a Service (MLaaS) platforms – with free trial possibility