We've designed a distributed system for sharing enormous datasets - for researchers, by researchers. The result is a scalable, secure, and fault-tolerant repository for data, with blazing fast download speeds.
Marvin is a deep learning framework designed first and foremost to be hackable. It is naively simple for fast prototyping, uses only basic C/C++, and only calls CUDA and cuDNN as dependencies.
F. Bogo, J. Romero, M. Loper, and M. Black. Proceedings of the 2014 IEEE Conference on Computer Vision and Pattern Recognition, page 3794--3801. Washington, DC, USA, IEEE Computer Society, (2014)
S. Levine, P. Pastor, A. Krizhevsky, and D. Quillen. (2016)cite arxiv:1603.02199Comment: This is an extended version of "Learning Hand-Eye Coordination for Robotic Grasping with Large-Scale Data Collection," ISER 2016. Draft modified to correct typo in Algorithm 1 and add a link to the publicly available dataset.