Huawei chips refine DeepSeek model in major leap for China’s AI self-reliance
Huawei has successfully used its advanced chips to refine the large 1.6 trillion parameter DeepSeek AI model. This marks a significant step for China's AI self-reliance, particularly in the complex area of model training, which has been a bottleneck compared to inference capabilities. The project, a collaboration involving multiple Shenzhen-based institutions and Huawei, demonstrated full-parameter post-training, updating the model's entire architecture. AI
IMPACT This advancement in domestic AI training capabilities could accelerate the development and deployment of advanced AI models within China, reducing reliance on foreign technology.