Put Tensorflow, Keras and MxNet Deeplearning Models on steroids

Accelerate your deep learning training times using multiple GPUs or experiment with multiple architectures in parallel using Tensorflow, Keras, and MXNet.

RocketML Distributed Deep Learning Suite is out!

The Artificial Intelligence is experiencing a renaissance. The current boom in AI is driven by the success of Deeplearning approaches.

But there is a Problem!

Elon Must funded @openai says deeplearning training times are increasing 5x the Moore’s law. This massive increase in training times are driven partially by growth in data sizes which is growing 2x the Moore’s law. Only way to keep up with data and training times is by resorting to Distributed approaches where RocketML specializes in.

RocketML Deeplearning suite distributes MXNet, TensorFlow, and Keras to multiple GPUs to speed up your training times. Let’s face it — any training times greater than 2 minutes is not fun. Engineers and Executives alike demand fast turn around times. Even as individuals we want everything fast. Somehow poor data scientists are made to wait for hours and sometimes days to run ‘one’ experiment while building deep learning models. We are on a mission to fix that challenge.

Following thought is extracted from Andrew Ng’s presentations about why HPC is so important for Machine Learning:

To build models with superior accuracy Data Scientists and Researchers need to iterate on their ideas. If a single iteration takes hours and days, training time becomes a deterrent to experimentation. Slow training is not an experience that is conducive to research and development of models. In light of increasing data sizes, flat computer performances, Distributed Machine Learning is an indispensable tool for researchers to get their job done.

Accelerate your overall model tuning time by experimenting with multiple architectures in parallel using RocketML distributed Deep Learning suite. It is fully built so that data scientists don’t have to worry about tuning their compute infrastructure and waste time on systems engineering tasks. They can instead focus on building better models, understanding their data, communicating their ideas and experimentation to management.

RocketML distributed Deep learning suite is available on AWS marketplace on a “pay as you go” service — where you don’t have to worry about long term contracts. Take it for a test drive! Contact me at vinay@rocketml.net or santi@rocketml.net if you like longer free trial duration or support. We are absolutely thrilled to hand hold and make you successful.