The Commands For Autonomous Vehicles (C4AV) challenge requires participants to solve an object referral task in a real-world setting. More specifically, we consider a scenario where a passenger can pass free-form natural language commands to a self-driving car. This problem is particularly challenging, as the language is much less constrained compared to existing benchmarks, and object references are often implicit. The challenge is based on the recent \texttt{Talk2Car} dataset. This document provides a technical overview of a model that we released to help participants get started in the competition. The code can be found at https://github.com/talk2car/Talk2Car.