Nervana Systems launched Nervana Cloud, a hosted hardware and software platform for deep learning that transforms how knowledge is extracted from data. It comes with a set of tools to enable small and large data science teams to build deep learning solutions that tackle complex data science problems, while building and scaling solutions on combined hardware and software cloud for machine intelligence.
Nervana Cloud enables organizations of all sizes to build and deploy deep learning solutions without investing in infrastructure or in a large machine learning / data science team. The company was founded in 2014 to make scalable artificial intelligence (AI) widely accessible and to enable disruptive new applications across a range of industry verticals.
Nervana Cloud enables healthcare, agriculture, finance, online services, automotive, and energy organizations to effectively use deep learning to solve their biggest data challenges.
With Nervana Cloud, data teams can now develop solutions to solve previously unsolvable problems such as increasing the accuracy of medical diagnoses, helping feed a growing global population through increased food production, reducing credit card fraud, improving financial returns, building intelligent cars, and conducting more efficient energy exploration.
Nervana Cloud, based on neon, Nervana’s open source deep learning framework, is a full-stack solution optimized to handle complex machine learning problems at scale. Neon is a deep learning framework available, and Nervana Cloud is more than ten times faster than any other cloud platform.
neon is Nervana’s Python-based deep learning library that offers ease of use while delivering high performance. Its features include support for commonly used models including convnets, RNNs, LSTMs, and autoencoders; integration with GPU kernel library; 3s/macrobatch (3072 images) on AlexNet on Titan X (Full run on 1 GPU ~ 32 hrs); basic automatic differentiation support; framework for visualization; and swappable hardware backends: write code once and deploy on CPUs, GPUs, or Nervana hardware.
New features include Kepler GPU support; new data loader and serialization formats; and expanded model zoo now featuring deep residual nets for image classification, fast-RCNN for object localization, C3D video action recognition.
Nervana Cloud removes the complexities associated with developing deep learning solutions, allowing data scientists to build, train, and deploy deep learning-based AI solutions using internal data. Nervana Cloud is available either through the public cloud or an on-premise hybrid cloud solution.