Yuanrong Qixing plans to develop VLA models based on NVIDIA Thor chips

Yuanrong Qixing plans to develop the VLA model (visual language action model) based on NVIDIA Thor chips, which is expected to be officially launched in 2025. The VLA model can connect vision, language, and action, and can identify and describe road environments, traffic signs, road participants, etc., and understand complex interactive events, hidden semantic information, and logical reasoning in traffic scenes.