Establishing a GPU Environment for AI Competition

YAI(Yonsei Artificial Intelligence)

Yonsei University’s Artificial Intelligence Society (YAI) has adopted and is utilizing Elice Project, one of the cloud services offered by Elice. In this interview, we sat down with Yoonjung Park, the Vice President of YAI, to discuss their implementation experience and in-depth reviews of using Elice Project.

Q. Please introduce yourself and YAI.

Hello, I am Yoonjung Park, the Vice President of the Yonsei University Artificial Intelligence Society, YAI. Established in 2018, YAI is Yonsei’s first AI society aimed at creating social value through artificial intelligence. We currently have around 70 members participating in various projects and research initiatives, driving our society forward.

Q. Quite a substantial number of members are involved. What led YAI to adopt Elice Project?

YAI hosts a biannual competition called ‘YAICON’. To efficiently conduct this competition, GPU usage is essential. Additionally, we run study groups where members read papers and implement models, activities that also require GPUs. Realizing the need for our own GPUs, we explored various options and decided on Elice Project due to its suitability.

Q. What was your first impression of Elice Project?

Our first impression was its ease of use. Although we often use terminal environments for server access, Elice Project’s intuitive environment, similar to Jupyter Notebooks, made it easy for anyone to download or upload files, which was impressive.

Q. Why did you choose Elice Project over other GPU services?

Other services typically have storage limits, incurring additional costs if exceeded, whereas Elice Project doesn’t have these constraints. The ease of project-based management was also a significant advantage. Moreover, Elice Project offers competitive pricing compared to international GPU services, making it cost-effective.

As a university society funded by grants, a flat monthly fee was more suitable than a usage-based billing model, further solidifying Elice Project as the optimal choice.

Q. How are you utilizing Elice Project, and what are your goals with its implementation?

As mentioned earlier, Elice Project is primarily used for training models and obtaining results during YAICON. It’s also actively used in our study groups and individual research.

Q. Was Elice Project user-friendly when you first used it?

Many users now favor VS Code, but for those making smaller investments or new to such tools, Jupyter Notebooks offer more convenience. While other GPU services had complex UI/UX, Elice Project’s design was straightforward and user-friendly. However, additional usage guides would be beneficial.

Q. Could you share any outcomes achieved using Elice Project?

Since Elice Project was mainly used for YAICON, we can showcase the winning projects from our competition. Below is an image of a project from the winning team.

Q. Do you find Elice Project pricing reasonable?

Considering the provision of A100 GPUs, the pricing is reasonable. While the society doesn’t necessarily need A100 GPUs, the performance, speed, and memory of high-performance GPUs are crucial, making Elice Project’s pricing for A100 GPUs a rational choice.

Q. Would you recommend Elice Project to others? If so, why?

Yes, I have recommended Elice Project to colleagues in other labs. Graduate students often need to work intensively near paper deadlines, and I encourage them to try Elice Project. The access to A100 GPUs is a significant advantage for students.

Q. Can you briefly outline YAI’s future direction?

As previously mentioned, YAI aims to create social value through AI and will continue to study and research with this vision. We also plan to hold the YAICON competition twice a year consistently.

Yonsei University’s Artificial Intelligence Society (YAI) has leveraged Elice Project to enhance the efficiency of their research and activities. The performance of A100 GPUs and reasonable pricing have been particularly beneficial. YAI will continue striving to create social value through AI. For those seeking high-performance GPUs for research and development, we recommend exploring Elice Project.

👉 Learn more about Elice Project

👉 Consult about Elice Cloud GPU Services

*This content is protected under copyright law and belongs to Elice.

*Unauthorized secondary processing and commercial use of this content are prohibited without prior consent.