With all the ICOs coming out lately, it is hard to get your target set on a real winner. Luckily companies such as TensorBit are making that a lot easier. TensorBit is introducing a concept to utilize all those GPUs that are mining “worthless” hashes. Instead TensorBit will utilize those GPU cycles to solve important machine learning and artificial intelligence computations. TensorBit will incentivize miners to solve these computations and offer this power as a service to potential companies such as Tesla, or Toyota which can greatly speed up their innovation of self-driving cars.
One of the founders of TensorBit says that “the company is at the forefront of two huge industries that are still in their infancy stages. That is, machine learning and AI along with blockchain technology. It makes a lot of sense. “ Right now TensorBit plans on releasing its initial coin offering in September where it hopes to raise funds to start building out its service.
This company could really be on to something huge as one of the major complaints of developing artificial intelligence is the time it takes to train the neural networks. As an example, using high priced GPUs offered by NVIDIA, can take hours, days or weeks to train against very large datasets or big data, a term that is frequently used in this day and age. TensorBit claims that it can reduce training times from weeks to days or hours to seconds. If this is true than TensorBit is truly on to something amazing.
An excerpt from their whitepaper visualizes some of the speed differences TensorBit can offer compared to the competition.
Server config: Dual Xeon E5-2699 2.6 ghz Performance measured on pre-production hardware. ResNet-50 Training on Microsoft Cognitive Toolkit for 90 Epochs with 1.28M imageNet dataset. A smaller bar is better.
TensorBit looks like a real winner and one worth looking at if you are wanting to invest into an ICO. This is a company where buying at the ground floor and holding out for a few years should certainly pay off in dividends.
Check Out These Popular Articles
Sorry. No data so far.