XPeng Motors Unveils End-to-End Large Model at AI Intelligent Driving Conference

TapTechNews May 20th news, during the XPeng AI Intelligent Driving Conference tonight, XPeng Motors released the so-called 'first mass-produced on-board in China' end-to-end large model, and the vehicle can directly output the control of the vehicle through the content input by the sensor, and the training content will also be upgraded from pictures to videos.

He Xiaopeng said that the goal of XPeng Motors' end-to-end large model is to iterate once every 2 days, and the intelligent driving ability will be increased by 30 times in 18 months. By the third quarter of 2024, it will be able to drive on every street in the country.

According to the introduction, XPeng's end-to-end large model will achieve a 2-fold increase in intelligent driving ability, a 2-fold increase in perception distance, and the ability to recognize more than 50 targets; it can recognize waiting areas, tidal lanes, and can read road sign texts and infer the intentions of traffic participants; in complex scenarios, the jerking, stalling, and safety takeovers are all reduced by nearly 50%, and a Level L4-like intelligent driving experience will be achieved in China in 2025.

The XPeng AI Intelligent Driving Conference is still ongoing, and TapTechNews will bring follow-up reports.