Enot.ai debuts answer to optimize deep neural networks

Read Time:2 Minute, 21 Second


Did you miss a session on the Information Summit? Watch On-Demand Right here.


Latvia headquartered Enot, an AI startup backed by New Nordic Ventures, has introduced the launch of a devoted framework to optimize deep neural networks.

The AI-driven providing, as the corporate explains, comes with a Python API that may be built-in inside numerous neural community coaching pipelines. It then automates the seek for optimum community structure, contemplating numerous {hardware} and software-centric parameters, together with RAM, latency, mannequin dimension constraints and operation kind.

“Enot.ai’s framework takes a educated neural community as enter, after which it selects a subnetwork that has the bottom latency and may guarantee no accuracy degradation,” the corporate informed VentureBeat. 

The entire course of makes the neural community sooner, smaller, and extra energy-efficient, fixing the challenges generally confronted by builders worldwide.

Enot’s Impression

Enot claims that its answer has the potential to assist AI builders and firms obtain as much as 20 occasions neural community acceleration and as much as 25 occasions community compression. The advantages even assist lower down computing {hardware} prices by as a lot as 70%.

“Enot is on the forefront of next-level AI optimization, serving to deliver quick, real-time ranges of AI development … Our journey has solely simply begun with examples such because the Weedbot laser weeding machine that gained 2.7 occasions acceleration, because of the Enot framework,” Sergey Aliamkin, CEO and founding father of the corporate, mentioned.

General, the corporate claims to have run pilot tasks with over 20 corporations, together with main gamers reminiscent of PicsArt, LG, Huawei, Dscribe and Hive.aero.

In a single case, it accelerated a picture enhancement neural community by 13.3 occasions for a smartphone producer. The optimization diminished the neural community depth from 16 to 11 and diminished the enter decision from 224 x 224 pixels to 96 x 96 pixels, with none lack of accuracy, the corporate mentioned. It additionally had one other challenge with the identical firm the place the framework delivered 5.1 occasions acceleration for a photograph denoising neural community, with none change in high quality.

“Earlier than assembly us, they already had a number of clients, together with giant worldwide tech corporations like LG, Huawei, Sony. That confirmed for us that Enot is fixing a business-critical difficulty within the neural community house that can not be solved internally nor are there any possible options obtainable available in the market,” Dmitry Saikovsky, Common Associate of New Nordic Ventures, informed Venturebeat. 

Different gamers trying to resolve the identical drawback are Deci.ai, OctoML, DeepCube, Deeplite, NeuralMagic, and DarwinAI.

VentureBeat’s mission is to be a digital city sq. for technical decision-makers to realize data about transformative enterprise expertise and transact. Be taught Extra



Supply hyperlink

Happy
Happy
0 %
Sad
Sad
0 %
Excited
Excited
0 %
Sleepy
Sleepy
0 %
Angry
Angry
0 %
Surprise
Surprise
0 %

Average Rating

5 Star
0%
4 Star
0%
3 Star
0%
2 Star
0%
1 Star
0%

Leave a Reply

Your email address will not be published.

Previous post 4 methods personalization will help manufacturers reconnect with customers
Next post When We Stop to Perceive the World assessment: Haunting and astonishing