What : Train machine learning models better and faster: with the use of Raven Protocol AI/ML engineers will be able train models better an faster.
Why : As faster model training will save time and cost and increase productivity eventually.
How : Conventional distribution methods have an inherent latency in their network. Large chunks of data need to be passed between machines. However Raven’s unique approach to distribution solves latency by chunking the data into really small pieces say bytes, maintaining its identity, and then distributing it across the host of devices with a call to action gradient calculations.
What: If you have a mobile device or a laptop or a desktop you can support Raven Protocol.
Why: When you will support Raven Protocol with your compute power you will get Raven token in return.
How: The Model is intact at the Master Node, and the heavy lifting is distributed in the tiniest snippets of data subsets over the network of contributors. The resultant gradients, after the calculations that happen at the node or contributor ends, are sent back to the Master Node.
What: Raven facilitates both Data and Model Parallelization approaches to form a different model of distribution.
Why: To make the system robust enough to handle the contributor addition or deletion, making the whole Raven training sustainable.
How: Raven uses dynamic computation graph not the static. A major difference between static and dynamic computation graph is that in the former, the model optimization is preset, and the data substitutes the placeholder tensors. Whereas, in the latter the nodes in a network are executed without a need for any placeholder tensors.
What: Raven Protocol is developed using Python and C++
Why: Because it is designed to keep no dependency on the architecture of the machine, nodes in the network can be as simple as a Javascript client.
How: For example a website could install a javascript snippet and execute gradient calculations on page load.
What: Raven Protocol aims to be the default framework for AI/ML training in near future.
Why: As beta customers of the Raven Protocol are very happy.
How: Customers will be paying raven more than a competitor company as raven delivers a solution to them that trains models faster.
For more details please visit below website:
