Caffe is a deep learning framework made with expression, speed, and modularity in mind. It is developed by the Berkeley AI Research Lab (BAIR) and community contributors.

It contains caffe executables, configured with CUDA.