Colorful, NVIDIA and Mars Era Education: How RTX AI accelerates the implementation of CG digital art creativity

On July 24, Colorful, NVIDIA and CG vocational education brand Mars Era held the inauguration ceremony of the “RTX AI Creation Acceleration Experience Zone”.

Ms. Cao Mengqi, ecological cooperation manager of NVIDIA Studio, Mr. Chen Chudong, product manager of Colorful Mobile System Division, and Mr. Wu Daji, director of AI design teaching and research of Mars Age Education, attended the event.

Mr. Wu Daji delivered a speech at the event. He mentioned that the widespread application of AI technology is promoting the continuous transformation of traditional education models and is also gradually changing the working methods of the CG design industry. Mars Era Education integrates AI technology into CG course design. And explore the integration of AI technology and digital art education.

Ms. Cao Mengqi introduced RTX GPU and AI technology to the audience. NVIDIA RTX GPU is equipped with a dedicated AI Tensor Core and provides RTX AI technology and tools to bring native acceleration capabilities of generative AI.

Mr. Chen Chudong introduced to you the new Colorful AI PC all-round notebook "Yuan N14", which is equipped with NVIDIA GeForce RTX 4070 laptop GPU, has a 17.9mm + 1.53kg body, and has a built-in "Hongguang AI" artificial intelligence assistant.

After the event, as the only specially invited media for this event, this website conducted in-depth interviews with three guests who attended the inauguration ceremony, focusing on the current development of the CG digital art education industry and the current status of NVIDIA Studio ecological construction.

Interview guest

△ Mr. Chen Chudong, product manager of Colorful Mobile System Division

△ Ms. Cao Mengqi, ecological cooperation manager of NVIDIA Studio

△ Mr. Wu Daji, AI design and research director of Mars Age Education

Treating AI as a tool is the basis for making good use of it

At the beginning of the interview, we asked Wu Daji his opinion on the issue of "whether AI will subvert the development of the CG industry". He believed: "AI can be regarded as a part of the CG workflow. Tools, you must use it to serve yourself. What really determines your level in the CG industry is always creativity and ideas." Wu Daji said frankly that many students have been troubled by this problem, but after learning, they will find that they have good ideas. Most importantly, because people can decide how the tool is used and the final result. The biggest role of AI at this stage is that it greatly improves the speed of commercial effect production and fundamentally solves the efficiency problem. This is also the significance of the establishment of the "RTX AI Creation Acceleration Experience Zone" by the three parties.

When asked about the impact of Colorful hardware and RTX AI technology on teaching scenarios, Wu Daji felt that embracing new software and hardware is necessary for the CG industry. Hardware is the foundation because the entire CG workflow requires Regardless of computers and graphics cards, AI technology is a means of improvement that helps simplify workflow and production cycles.

He gave an example in the "Compulsory Course for Commercialization of AI Design" video series, using SD to run SDXL to generate 100 pictures with a resolution of 1024*1024. A powerful NVIDIA GeForce RTX 40 series graphics card can already make the entire The generation time is compressed to about 500 seconds, with an average of 1 picture in 5 seconds. Coupled with TensorRT acceleration technology, the time can be further reduced by a factor of 1. Powerful hardware can greatly improve teaching and learning efficiency. Software and hardware selection is very important.

In addition to providing students with good hardware to experience, what teachers have to do is to adjust the teaching direction in a timely manner and integrate AI content into the courses. Mars Era currently covers "AI + graphic design" and "AI + original painting design" Key CG skills courses such as "AI + 3D Modeling" and "AI + Short Video" are designed to help students quickly adapt to the changing commercial environment and comply with the future employment trends of CG industry talents.

The efficiency of combining software and hardware is maximized, and the upper limit of creation depends on the person

In this event, the "RTX AI Creation Acceleration Experience Zone" at the event mainly demonstrated SD ComfyUI + InstantID Superman photo generation, SD ComfyUI + LivePortrait digital human adaptation Equipped with facial expression animation effects, NVIDIA Canvas software graffiti conversion to real-life drawing and other demonstration examples.

With the help of NVIDIA RTX technology, it only takes about 10s to generate 6 different styles of Superman pictures, and the longest does not exceed 30s. The expression animation-driven demonstration can basically be completed within 15 seconds, and the response speed of Canvas' live scene transfer is almost real-time. Mengqi Cao explained in detail the role of RTX GPUs in these AI workflows.

"Hardware computing power is the most fundamental. For example, the computing power of 40 series GPUs is above 200TOPS, and Stable Diffusion runs on the CUDA ecosystem and supports TensorRT acceleration, which can maximize computing and inference efficiency. SD ComfyUI, WebUI These have a high threshold for getting started, and you need to learn to adapt and quickly deploy a highly reproducible workflow; NVIDIA Canvas provides a lower threshold for getting started, which depends on your understanding of drawing. Good aesthetics and artistic thinking will increase infinitely. The upper limit of works.

AI ecological closed loop requires the gathering of wisdom from many parties

As for why PC hardware partners chose Colorful, Wu Daji mentioned that when they first cooperated in 2022, Colorful just launched the Colorful ProMaster A1 designer host, which is specially designed for Designed by a group of designers, hardware performance will greatly speed up teaching in the Mars era. Chen Chudong said that Colorful has always paid attention to specific scenarios that require hardware implementation in AI applications, such as three-dimensional animation and CG. Cao Mengqi added: " NVIDIA is still far away from end consumers. Colorful and Mars Era are both partners and an important bridge to evaluate the market. They can sense consumers' feedback on products and needs in a more timely manner. Cooperation is not a simple combination of software and hardware, but a discussion. On many industry issues, we have complementary opinions. ”

Of course, the premise of the tripartite cooperation is to achieve consensus. We hope to fully tap the potential of RTX AI hardware performance and software ecology through the field of CG digital art education, and promote the innovation and development of AI technology in CG education and all walks of life. The cooperation process There is running-in and growth, and it must be a win-win situation in the end.

Interview transcript (condensed version)

Q1

This site: How did you define the core characteristics of Colorful New Source N14 AI PC when designing it? Chen Chudong: For this product, the first thing we pursue is an all-round experience. When designing it, we mainly consider the type of users who want both high performance and a thin and light body. So how can we achieve it? With a weight of less than 1.5kg and a thickness of 17.9mm, the i9-13900H processor and RTX 4070 graphics card are packed into the body. The performance release of the whole machine when double-baked reaches 100W, which is first-tier among laptops of the same size. .

Secondly, we hope that users can bring it and use it in various application scenarios, such as some high-load scenarios, such as 3D rendering, video editing, and scenarios that require high AI computing power. Small body, high performance, portable use, taking into account thinness and high performance, these are the advantages of Yuan N14 hardware configuration

Then the user experience of Yuan N14, we chose an all-metal body, which is very comfortable to touch. ; The surface of the fuselage is made of CNC technology, which looks more textured. We have done some research on users in terms of color matching. From the overall data, silver is the most accepted by users, so the base model of Yuan N14 is made of silver and blue. The main color is to give users more choices. The source N14 is now available in 14 inches, and we will update other sizes and colors of the same type in the future. In short, we hope to give users the right to choose in terms of design and experience. We hope to be comprehensive.

Q2

This site: Mars Age will fully integrate AI into its full range of course products in early 2023. Will Colorful, as a partner, carry out other cooperation with Mars Age in education?

Chen Chudong? : Nowadays, most AI models have been integrated into our lives. Many AIs are used in various fields such as education, technology, and media creation. Improving efficiency is the core value brought by AI at this stage

For example, this time Colorful. , NVIDIA Studio cooperates with Mars Era in the RTX AI Creation Acceleration Experience Zone. We can see that future AI applications include many scenarios that require hardware implementation, such as 3D animation and CG film and television in the field of education. It can all be implemented through hardware. In usage scenarios, it allows creators and students to achieve real efficiency improvements in their work and life.

Colorful product hardware is then paired with NVIDIA’s software drivers, as well as the Mars Era’s accumulation in the education field and supporting teaching experience. , these things are combined together to help creators use them more efficiently in specific scenes, and spend more energy on creation instead of simply moving bricks. This is the original intention of our cooperation.

In future cooperation, we will continue to maintain this original intention, provide our users with more efficient products, and hope to help them create "efficiency value."

Q3

This site: Before coming here, I watched some videos on Mars Era (a required course for AI design commercialization). One of the videos mentioned the idea that the traditional design workflow in the CG industry will be subverted by AI, because PC hardware with AI computing power can greatly simplify the creative process and improve creative efficiency. In the past, it took several days to model, render, and color the generated pictures. Now, AI tools can be used to produce pictures in 10 minutes, and then it takes half a day to modify them. Just BUG. The factors that restrict creators are no longer design, painting and software operation abilities, but ideas, aesthetics and computing power.

So I am very curious, what impact will the introduction of NVIDIA Studio and Colorful AI PC hardware have on the educational scene in the Mars era? Will teachers in the future spend more energy cultivating students’ creative design thinking?

Wu Daji: Your questions are quite typical. In fact, many people will have some misunderstandings about AI, such as the subversion you just mentioned. From the current perspective, this view is not comprehensive enough. AI is essentially a tool in the workflow for the CG industry or us. It can speed up the creative process. Maybe in one link, it can greatly accelerate and improve efficiency. A certain AI tool can greatly improve the effect production of our entire commercial project. It is very helpful, but when looking at the industry, we must continue to observe and summarize what the future development of the industry will be like.

In other words, if you want to do commercial projects, students still need to systematically learn and train in traditional CG related knowledge. What we teach students is how to complete a purely commercial work from beginning to end. It is not only theoretical knowledge, but also our experience and thinking about the CG industry. After all, it is people who do the project, not AI. As a tool attribute in workflow, AI plays more of an auxiliary role.

In the CG industry, everyone is actually very keen to embrace new AI hardware and technology. After all, our work is done through computers, and hardware and technology are also what we eat. Because computer hardware and NVIDIA graphics card technology are of great help to our entire workflow, through cooperation with NVIDIA Studio, we will also explore the improvement and optimization of work content based on the role of AI, and continue to fine-tune the teaching process. Improve the practicality of teaching.

On the other hand, we have also been working hard to introduce advanced hardware equipment into the classroom so that students can experience it for the first time. We will even accept some commercial projects for a fee, let students work in groups to complete, and then send all the income to For them, everyone will be more motivated to learn after receiving the red envelope, and they can also accumulate valuable business experience during the learning stage. Some students have projects that have been launched on TV stations or movie theaters before they graduate. Ultimately, we hope to help students adapt to complex commercial environments more quickly and efficiently.

Q4

This site: Mars Age Education RTX AI Creation Acceleration Experience Zone This cooperation, how did Colorful, NVIDIA and Mars Age hit it off in the first place? Can you share the story behind it?

Wu Daji: In 2022, NVIDIA Studio launched their Canvas creative drawing tool. When we saw this tool, we felt that the AI ​​technology behind it might have an important impact on our future teaching and industry. So we established cooperation with NVIDIA Studio and together we built the first NVIDIA Canvas comprehensive practice base in China. Another opportunity was that Colorful launched a PC console specifically targeted at designers, the Colorful ProMaster A1 designer console. It was precisely because of this opportunity that the GeForce RTX creative design space was born.

Later, after case creation testing, Canvas has great practical prospects for Matte Painting (digital painting) in the film and television industry, so we also integrated this software into the course. We were very happy with the cooperation process. In fact, everyone Very satisfied with the entire cooperation process and results. With this foundation in place, we will conduct another round of cooperation in 2023. This time it is a commercial creation open class cooperation in the direction of pure AI. It is also an exploration for everyone to do together.

We (Mars Age Education), NVIDIA Studio, and Colorful shared together, and later we had some more in-depth cooperation. For example, NVIDIA launched the TensorRT acceleration plug-in for Stable Diffusion. We collaborated with NVIDIA Studio to produce the "NVIDIA TensorRT Stable Diffusion Creation Acceleration Guide" white paper, which is a very important attempt by Mars and NVIDIA to commercialize AIGC teaching this year.

Of course, as I mentioned before, the CG industry will naturally pay attention to and embrace some new technologies and new equipment, and will have in-depth discussions with partners on how the industry will develop and change in the future, and how to update workflows and talents. These are the needs that we will continue to explore. As early as 30 years ago, when the Chinese CG industry just emerged, we entered this industry. It can be said that we have enlightened and led the development of the CG industry along the way. Therefore, from the perspective of responsibility, everyone also wants to do something for this industry. After all, our Students are the backbone of the future industry.

Q5

This site: I noticed that the RTX AI Creation Acceleration Experience Zone can be said to be a fusion of hardware, technology and ecology among the three partners. What role do you think NVIDIA plays in it?

Cao Mengqi: From NVIDIA’s perspective, the RTX AI Creation Acceleration Experience Zone we launched this cooperation hopes to bring some advanced AI workflows to students through the RTX AI PC brought together with Colorful Acceleration and boost. NVIDIA's RTX GPU uses computing power to provide computing power support, optimization and acceleration for all similar AI workflows, which is the role of the source of power.

Nehmen Sie Stable Diffusion als Beispiel: Die gesamte SD läuft auf dem CUDA-Ökosystem von NVIDIA, sodass unsere Grafikkarte ihre Leistung auf natürliche Weise beschleunigen kann. Dies ist untrennbar mit der Engine und der Technologie verbunden, die wir verwenden. Darüber hinaus erfreut sich das Konzept des „KI-PCs“ mittlerweile in der PC-Branche großer Beliebtheit. Jeder hofft, dass er mit aktuellen und zukünftigen KI-Tools problemlos umgehen kann.

Tatsächlich verfügt NVIDIA über eine Reihe von Standards zur Messung der KI-PC-Leistung. Im Allgemeinen wird sie in der Einheit TOPS gemessen, die angibt, wie viele Billionen KI-Operationen die GPU pro Sekunde ausführen kann. desto besser ist die Leistung des Computers für KI-Computing. Derzeit liegt die TOPS-Rechenleistung der Grafikkarten der NVIDIA RTX 40-Serie grundsätzlich über 200, was ein guter Prozessor für Szenarien sein kann, in denen Verbraucher in ihrem täglichen Leben lokalem KI-Computing ausgesetzt sind, wie etwa KI-Konferenzen, Programmierfelder usw Mehr oben erwähnt. Erweiterter KI-Anwendungsworkflow.

Apropos unsere Canvas-Software für kreatives Zeichnen: NVIDIA Canvas ist eine Software für Designer, die schnell verwendet und einfach erstellt werden kann. Sie soll dafür sorgen, dass Kreative jeden Alters und Berufes Spaß am Schaffen haben. Unsere Idee ist sehr einfach: Mithilfe der KI-Technologie vermitteln wir Ihnen das Gefühl, dass die Erstellung tatsächlich sehr einfach werden kann.

Zuerst dachte ich, Canvas sei ein Unterhaltungstool, aber durch unsere Zusammenarbeit mit Mars Age sahen wir, wie Profis ihr kreatives Denken einbrachten, um viele unvorstellbare Anwendungsszenarien für Canvas zu inspirieren. Als unsere eigenen technischen Kollegen Canvas vorführten, haben sie vielleicht Dinge gezeichnet, die wie einfache Zeichnungen aussahen, aber nachdem sie es Lehrern im Mars-Zeitalter zum Ausprobieren gegeben hatten, war der Endeffekt sehr schockierend und kann nur von Profis durchgeführt werden . Eine professionelle Sache. Die Werke, die sie mit Canvas geschaffen haben, unterscheiden sich von der Komposition bis zum ästhetischen Stil völlig von unseren.

Q6

Diese Seite: Als Demonstrationskooperationsprojekt im Bildungsmodellraum von RTX AI beschleunigte die Erstellung, ob die Lehrerfahrung der Mars-Ära und die Hardware-Überlegungen von Colorful auch NVIDIA Studio dazu beitragen werden, gezieltere Anstrengungen in der Bildung zu unternehmen Ökologie-Innovationen, wie funktionale Feature-Updates von Studio-Plattformanwendungen und -Tools, Software-Workflow-Optimierung usw.?

Cao Mengqi: Ich denke, dass es auf jeden Fall passieren wird. Es handelte sich ursprünglich um eine auf NVIDIA Canvas basierende KI-Zeichenanwendung Zusammenarbeit Während dieser Zeit haben wir von Lehrern und Schülern in Mars Era viel Feedback zu Canvas erhalten, darunter auch einige Verbesserungsvorschläge.

Zum Beispiel kann die Ausgabeauflösung von Canvas zunächst nur eine niedrige Auflösung von 512*512 erzeugen, aber wir haben später Verbesserungen vorgenommen. Wir werden das ausführliche und detaillierte Nachfrage-Feedback der Partner in tatsächlichen Anwendungen besprechen und es in zukünftige Produkt-Upgrades und Iterationspläne integrieren, um es schrittweise zu verbessern. Beispielsweise ist der 3D-Rendering-Modus, der später eingeführt wird, auch das Ergebnis der Kollision aller Köpfe, daher schätzen wir jede Zusammenarbeit mit Mars Age.

Was die RTX AI Creation Acceleration Experience Zone betrifft, denke ich, dass sie eine gute Gelegenheit für NVIDIA ist, die tatsächliche Anwendung von KI in der Designbranche zu verstehen, einschließlich der Möglichkeiten normaler Benutzer mit NVIDIA-Grafikkarten und ihrer Anforderungen an Hardwarepunkte und Wie Sie die Effizienz auf der Softwareseite verbessern möchten, das ist alles wichtig.

Wir haben normalerweise eine gewisse Distanz zu Endverbrauchern und Designergruppen, um einige Bedürfnisse und Rückmeldungen schnell zu erfassen. Dazu benötigen sie auch Stable Acceleration-Plugins Sie haben eine bessere Lern- und Schöpfungseffizienzsteigerung.

Ich glaube, dass wir durch unsere Reihe von Dreierkooperationen auch in Zukunft die Entwicklung der gesamten Designbranche auf dem Weg der KI vorantreiben und fördern werden.

The above is the detailed content of Colorful, NVIDIA and Mars Era Education: How RTX AI accelerates the implementation of CG digital art creativity. For more information, please follow other related articles on the PHP Chinese website!