search

trust-ai / SafeBench

A Benchmark for Evaluating Autonomous Vehicles in Safety-critical Scenarios
https://safebench.github.io
MIT License
96 stars 21 forks source link

can you supply the inference code? #2

Closed wangchangquan closed 1 year ago

wangchangquan commented 1 year ago

1，Whether the reinforcement learning algorithm mentioned in the article needs to use the data generated by Safebench for training？ 2，Can you provide inference code and pre training model ？such as DQN and DDPG

garyxcj commented 1 year ago

Hi,

Thank you for your interest in our work!

  1. The RL models used in our paper are trained with data in SafeBench. However, you can test your models trained with any other data. They don't necessarily need to be trained with SafeBench.
  2. You can download the models used in our paper here. We also updated the README. Please follow the instructions in step 4 to evaluate the models.

Please let me know if you have further questions!