Code Dreams: A Journey from Obscurity to Success
Code Dreams: A Journey from Obscurity to Success
In the hustle and bustle of the city, people chase after their livelihoods and dreams. Among them is Zhou Yifeng, an independent developer and MindSpore mentor. Actively participating in the open source community, he is skilled in compiling code, locating bugs, and guiding inexperienced developers. Through tremendous hard work, he has won many awards in related competitions. But his success was not a piece of cake. It was Zhou's struggle with life and persistence in pursuing his dreams that transformed him from obscurity into a successful code developer.

Rising from Obscurity
"I'm an independent developer, but also runs errands to make ends meet." Zhou says.
His journey to becoming a developer started in secondary school when he discovered programming on an app and became interested in it. He decided to major in software engineering and worked hard to learn the necessary skills, and participated in various programming competitions.
However, due to poor health and other circumstances, he dropped out of the secondary school and later attended a vocational school. Despite his impressive programming skills and awards from provincial and national competitions, Zhou struggled to find stable employment due to his lack of formal education and physical limitations. After graduating from the vocational school, he tried to find a job at an enterprise but failed.
To support himself, he began running errands. But his passion for programming never faded, and it continued to inspire him like a small flame illuminating his path forward. In his free time, he continued to learn and improve his technological knowledge.
Four years ago, Zhou discovered deep learning and was immediately captivated. However, learning it was no easy feat as it required both professional knowledge and hardware resources. Despite the lack of incentives and recognition, Zhou persevered in his studies and soon discovered the MindSpore open source community.
In June 2020, he participated in a model porting competition based on MindSpore papers and successfully ported a model for the first time. From then on, he became a fan of MindSpore due to its usability, even though it had only been released for less than three months at the time.
"MindSpore has many new features, and I prefer its support for both dynamic and static graph modes." Zhou says in excitement, "After joining the MindSpore community, I was given many opportunities to use its hardware resources. MindSpore supports not only GPUs but also Ascend AI Processors. During the competition, I used the Ascend 910 AI Processor, which was faster at training the ported DenseNet model than the V100 GPU. And at the time, the A100 GPU had not even been launched yet. Most importantly, I have a lot of opportunities to communicate with and learn from official staff in the MindSpore community. This experience I rarely had in communities of foreign frameworks."
Despite the challenges and setbacks he has faced, Zhou remains determined to continue learning about technologies and participating in MindSpore and other communities in pursuit of his dream.
Keeping Going to Realize His Dream
Zhou continues to run errands to support himself, but he now devotes more time to learning professional technologies and participating in AI competitions and activities. "Though I'm an errand boy, I place a high value on professional skills. By taking part in competitions like those organized by MindSpore, I not only improve my abilities but also have the chance to win prizes and bonuses. This helps to ease the financial burden of life and allows me to spend less time running errands."
But why does Zhou Yifeng choose MindSpore?
He says that compared with other open source communities, the MindSpore community is more active and frequently organizes activities and competitions with bonuses or assigns tasks to attract and nurture developers.
Additionally, the community has provided Zhou with significant support, including access to computing power and technical assistance. Furthermore, the community has established different SIGs to facilitate communication and learning among developers based on their specific interests and concerns.
As he learns from the MindSpore community, Zhou also has made significant contributions, including fixing bugs, submitting pull requests (PRs) and open source code, and mentoring beginners. In 2022, he even served as a MindSpore mentor for the Open Source Promotion Plan (OSPP), guiding students in completing their projects.
"This was a valuable opportunity for both my students and me, and it also benefited the open source community," he says. Understanding the challenges of balancing technology learning with making a living, Zhou is committed to helping others.
Joining the MindSpore community has opened up new possibilities for him. As an accomplished AI developer, he can now realize his potential and look forward to a bright future instead of just doing odd jobs.
Power of Persistence
Persistence is normally the beginning of success. Over the past three years, he has ported and reproduced various models, including DenseNet, PDarts, FishNet, OSVOS, EfficientNetV2-S, AutoDis, MMoE, BiT, and Ultra-Fast-Lane-Detection. He has adapted these models for use with GPUs or Ascend and made their code available through the OpenI community, Huawei Cloud AI Gallery, or the models repository. Zhou has also assisted with the efficient inference and deployment of MindSpore models in the Ascend environment, such as RetinaFace for object detection and YOLOv4 for pet recognition. He even attempted to deploy RetinaFace in a CPU environment, though it was successfully deployed in the MindSpore foundation model.
Despite facing challenges caused by his ill health, Zhou remains determined. "As long as I focus on AI deep learning and address problems while insisting on what I can do, I believe that I can gain more in the future," he says.
From a delivery boy to promising developer, now Zhou Yifeng is one of 377 official mentors recognized by the MindSpore community.
Growing with Developers
At the 2023 AI Framework Ecosystem Summit, Ding Cheng, director of the MindSpore open source community, emphasized the importance of talent support for the growth of China’s AI industry. He stated that the AI framework is a crucial tool for developers and that MindSpore is dedicated to cultivating AI talent.
Since its inception, the MindSpore community has been committed to these goals and has developed a growth path for developers. In recognition of the importance of developers and technologies, MindSpore released China's first developer and mentor certification system for the AI open source community.
The MindSpore community offers a comprehensive program to support developers at different stages of their growth. In the entry-level phase, the community provides tutorials and courses, and partners with universities to introduce students to MindSpore. In the practice phase, developers can improve their skills through internships, projects, and competitions. In the research and innovation phase, the community offers academic incentives and computing power support to encourage original scientific research.
MindSpore has established partnerships with over 210 universities and research institutes worldwide, including all C9 universities. Through these partnerships, MindSpore has trained 2,500 teachers and over 200,000 students. In the area of scientific research and innovation, MindSpore has collaborated with more than 290 top research teams globally and has published over 900 MindSpore-based papers at AI conferences, ranking first in China and second in the world.
As an open and inclusive AI community, MindSpore provides a foundation for innovation for developers, who in turn support and expand the community. By the end of May 2023, the MindSpore community had over 4.74 million downloads, 13,000 contributors, 5,500 enterprises served, and 400 open source models.
The MindSpore community is considered one of the most innovative and dynamic AI open source communities, attracting a growing number of developers to learn cutting-edge AI technologies and pursue their dreams.