Lu ZHANG

@zhaone Hi, have you found the reason? I just tried to train the network using the default settings, and I also found the training is around twice as slow as...

@qinwenchuang Thanks for your comment. We will check it carefully and fix it if it's a bug.

@IngaVS Sorry for the late reply. The driving corridor of the SSC is consecutive axis-aligned cubes, so the inequality constraint is straightforward. For one dimension (speed profile) optimization, the trapezoid...

Well... I met with this problem and solved it by chance. The reason was about the versions of Eigen and Ceres in my system. So I solved it by following...

@scott-ettinger Hi, Thanks for your response! Actually, we are investigating whether the dataset is suitable for other tasks, such as traffic analysis in a fixed region (which is definitely beyond...

Hi @wqi, I find that the number of scenarios in each split (for the current version of dataset) is not the same as described in the paper: test 24984 |...

> The updated number of scenarios won't cause any issues with the test set evaluation on EvalAI. The evaluation server and computed results have been adjusted on the backend to...

@dengnaitian We implement a lightweight simulator based on ROS. The simulator uses the kinematic bicycle model for vehicle state update and it also has some simple methods to manage map...

@yukkysaito Hi, thanks for asking! We are currently dedicating to the real-world experiment and could not release a major update in the recent days. We hope we can public our...

@purewater0901 Hi. Thanks for your question! The current released version does not contain CFB and the belief updater (actually an intention predictor) since we are still cleaning the code. We...