Xiaomi Publicly Streams MiMo-V2.6 Reinforcement Learning Training
KuCoinFlashShare
Xiaomi announced a project by live-streaming the reinforcement learning training of its MiMo-V2.6 model. Led by Luo Fuli, the team has been focused on scaling reinforcement learning for nearly six months. V2.6 is still under training, with open-sourcing expected soon. The process uses 1,568 prompts per step, generating 16 rollouts each, totaling 2 billion tokens. Tasks including code, general, visual, cybersecurity, and chat agents are trained simultaneously. On-chain observers can track metrics such as task composition and training duration in real time. UC Berkeley’s Yichuan Wang noted that over 40,000 sandboxes are active simultaneously.
Source:Show original
Disclaimer: The information on this page may have been obtained from third parties and does not necessarily reflect the views or opinions of KuCoin. This content is provided for general informational purposes only, without any representation or warranty of any kind, nor shall it be construed as financial or investment advice. KuCoin shall not be liable for any errors or omissions, or for any outcomes resulting from the use of this information.
Investments in digital assets can be risky. Please carefully evaluate the risks of a product and your risk tolerance based on your own financial circumstances. For more information, please refer to our Terms of Use and Risk Disclosure.