Home > News content

After Tianyuan open source, Kuang also showed a "skill ": the launch of the Brain commercial version

via:博客园     time:2020/10/9 18:39:19     readed:677

Bian CE originated from Aofei Temple

Qubit report official account number QbitAI

2020 is the first year of AI landing has no doubt. In April, the National Development and Reform Commission identified artificial intelligence as

Sharp tools make good work. Domestic AI software has been ready to try.

In the first half of this year, Huawei, Kuangshi, Tsinghua and other universities and companies successively open source and self-developed AI frameworks, taking the first step of large-scale implementation of AI.

The domestic AI productivity platform started earlier: Baidu has

Can AI be built easily and quickly with a platform? Not really.

Looking at the current AI platforms, it is not difficult to find that these platforms either serve professional customers or provide ready-made solutions directly to enterprises with little development capability.

However, as the scale of the AI's business landing becomes larger and more diversified, the demand of enterprises will inevitably meet

They don't have very strong development capabilities, and there are no ready-made solutions available in the market, in a form

Such enterprises often encounter the following problems in AI commercialization:

  1. High technical threshold: professional talents are scarce, and the technical work involved is complex and difficult

  2. Capability fragmentation: lack of whole process tools, need to integrate and develop on the basis of scattered tools

  3. Long tailed requirements: scenarios are scattered and a large number of customized AI capabilities are required

  4. Safety

With a strong AI R & D capability and extensive AI commercial experience, a company is willing to

Come on. The latest news, Kuang now released for corporate customers

First commercialized

People familiar with open vision should know that Brain is a magic weapon for many years to start a business in AI field.

Brain Business Edition is a product-based integration of past Brain development processes, tools, and capabilities.

In 2011, AI technology represented by deep learning began to break out, and kuangzhi was founded in that year.

Because the industry does not have ready-made, scale-compliant AI ability to develop tools, open their own from scratch to design tools, build processes, sedimentation capacity, such work has been six years since 2014. As a result of these efforts, the Brain platform is used by all staff

With the ability of Brain, open vision to produce large-scale AI algorithms, it stands out in AI commercial scenes such as computer vision. Brain can be said to have passed the test of time and market, and open-minded itself is a successful case of Brain commercialization.

After six years of iterative upgrades, Open Vision is now exploring the commercialization capabilities of the Brain, releasing business customers

Broadview's Brain commercial version has three advantages:

1、 Through the value verification of more than 100 kinds of products of Kuangshi for many years, more than 1400 R & D personnel of Kuangshi are using it;

Second, has the strong engineering ability: Brain has integrated the open vision rich AI development experience and the leading technical tool;

3、 To provide customers with the whole process of supporting services, can be one-stop enterprise worries.

The biggest difference with other AI productivity platforms is that Brain does not have a platform before landing, but grows with the open-view commercialization process, the feasibility of which has been verified over the past six years and is born commercialized

As an AI start-up company, Kuangshi understands how AI platform can serve enterprises better.

How does Brain business edition diffe

As we all know, the three pillars of AI are data, algorithm and computing power.

Nearly all mainstream AI platforms consist of these three parts, Brain no exception.

Brain integrates MegData、 MegEngine for data management for AI model development and MegCompute. for computational power management

But open-sighted Brain commercial version is not mechanical integration, the direct use of three modules does not conform to business thinking habits.

For general AI productivity platform, enterprises need to use different products to develop and use AI algorithms, manage data, and schedule computing power, resulting in capacity fragmentation.

Brain commercial version this time

While other AI platforms, deep learning frameworks are providing code base or installation packages, Brain commercial version is more like AI algorithm development

The algorithm layer is used to solve practical business problems, the platform layer provides capability support, and the delivery layer ensures that the platform can be landed and the algorithm can be applied.


The algorithm of the algorithm, the hardware of the hardware. The Brain commercial version not only provides software to customers, but also offers privatization deployment schemes. Besides the products, it also provides a set of service systems, such as statutory development and platform use training, which can be described as enterprise end-to-end solution AI business landing

As for software, Brain commercial version covers the whole process of data management, model development, computing power scheduling and so on.

Besides algorithm software development platform, Brain commercial version can also provide customers with hardware delivery including cluster building and deployment.

This is very different from other AI platforms. While solving the problem of AI software development, enterprises no longer have to search for AI hardware providers and software and hardware adaptation problems.

Go to the enterprise

The upgrade of product logic is nothing more than to better serve customers.

Brain business edition originally came from the AI enterprise, now wants to reintegrate and then goes to the tens of thousands of enterprises.

In the past, if enterprises want to use AI to improve production process and production efficiency, they often need to recruit talents from 0 to 1, build teams, invest high cost and go through a long R & D cycle.

Broad vision of the Brain commercial version is in such an environment came into being.

Brain commercial version not only makes the AI platform more suitable for enterprises in the software logic level, but also has incomparable advantages in the actual process of dealing with enterprises.

first, Brain commercial version provides users with the hardware needed to deploy the AI model. As a result of hardware services, the Brain commercial version certainly supports privatization deployments and addresses enterprises

Secondly, the Brain commercial version provides the whole process supporting services, from the early program consultation, to the later deployment, operation and maintenance, training and so on, can be one-stop to avoid the worries of enterprises.

Both advantages accelerate the efficiency of landing AI. Broad view shows that the Brain commercial version can effectively shorten the time of 80% algorithm from demand to landing, and reduce the overall production cost of 55% algorithm.

Whether you want to build a new AI business, or AI existing R & D projects to improve R & D efficiency, you can use the Brain business version.

A product release is only the first step forward, and in the future Brain the commercial version can achieve three goals:

1. Support 1000 traditional enterprises to build their own AI capabilities and realize intelligent transformation;

2. Promote the implementation of 10000 AI long tail algorithms and improve AI application penetration rate;

Commenting on the value of the launch of the Brain commercial edition, Tanaka, senior technical director of the Open Vision Institute, said:

Due to the high technical threshold of deep learning and the scarcity of professional talents, it is difficult for traditional industries to attract suitable talents to build this ability in depth for a long time. As an AI enterprise, we believe that we have the ability and responsibility to lower the threshold for enterprises to deploy AI, help more enterprises use AI, make good use of AI, and benefit the public with AI.

This year, AI commercialization has entered deep water area. How to expand AI commercial scale is a problem that every enterprise is thinking about,

As the best known

An open source Tianyuan is the first step in building AI ecology and is now looking to commercialize its technology

China IT News APP

Download China IT News APP

Please rate this news

The average score will be displayed after you score.

Post comment

Do not see clearly? Click for a new code.

User comments