Rainwater Technologies reported on June 26 that at the Xiaomi Renchejia Full Emanagical Press conference, Lei Jun introduced the fact that the entire Xiaomi YU7 series is equipped with end-to-end assisted driving, and that the full hardware is standard across the range.
Specifically, the entire Xiaomi YU7 series is equipped with an NVIDIA THOR chip with 700 compute power, LiDar, 4d millimeter-wave radar, and more, and the vehicle also improves its advanced information sensing capabilities. This means that Xiaomi has used high specifications in its perception hardware, laying the foundation for implementing advanced assisted driving capabilities.
Xiaomi End-to-End Assisted Driving currently uses 10 million edits (edits are scene edits used to learn end-to-end large models). The volume of 10 million edits demonstrates Xiaomi’s investment in assisted driving data collection and training. The more data the model can learn and adapt to a variety of complex driving scenarios, the more reliable and safe assisted driving becomes.
Ten million clips won’t be final; subsequent learning clips are still being added. The scenario edits presented in the press release are more than three times as many as the previous version. This shows that Xiaomi is constantly using real road data to optimize its assisted driving system.
The vehicle’s control and deceleration control have been greatly improved, and the entire driving experience will be better. In addition, the vehicle recognizes obstacles more accurately and the vehicle can easily go around them. This means that the HelpDrive system greatly improves control accuracy and safety. Flexible acceleration and deceleration control improves passenger comfort, while more accurate obstacle recognition reduces the risk of accidents.
Under the program, the Xiaomi Yu7 will carry 10 million end-to-end assisted driver clips across the range, with the SU7 PR, SU7 Max and Su7 Ultra expected to be officially upgraded in late July. This news is certainly good news for Xiaomi SU7 owners, who will also get a chance to experience the latest end-to-end assisted driving technology. End-to-end assisted driving technology does not allow rules to be defined manually, allowing the vehicle to learn independently through big data training and adapt to complex driving environments that are theoretically closer to those of a human driver.
>
These are the most important things you can do for your organization.
Presence
Presence