search

yixiao1 / CILv2_multiview

This is a repository for paper Sequential Attention Learning for End-to-end Driving
Other
14 stars 2 forks source link

Is there a script for data collection from Roach RL expert? #3

Open donghaiwang opened 1 month ago

donghaiwang commented 1 month ago

Carla keeps crashing when I run Roach RL expert. Can you share the process of training and evaluating Roach RL expert? Thank you.

donghaiwang commented 4 weeks ago

Where are the evaluation results saved by the save_dict function in the CILv2_multiview\run_CARLA_driving\driving\utils\checkpoint_tools.py script? And how to get metrics in your paper using script, including SR, S.SR, T.L, Avg.RC, Avg.DS, C.V, C.L, T.L, O.L, R.Dev. Think you. @yixiao1

╒═══════════════════════╤═════════╤═════════╕ │ Criterion │ Result │ Value │ ├───────────────────────┼─────────┼─────────┤ │ RouteCompletionTest │ SUCCESS │ 100 % │ ├───────────────────────┼─────────┼─────────┤ │ OutsideRouteLanesTest │ FAILURE │ 0.64 % │ ├───────────────────────┼─────────┼─────────┤ │ CollisionTest │ SUCCESS │ 0 times │ ├───────────────────────┼─────────┼─────────┤ │ RunningRedLightTest │ FAILURE │ 3 times │ ├───────────────────────┼─────────┼─────────┤ │ InRouteTest │ SUCCESS │ │ ├───────────────────────┼─────────┼─────────┤ │ AgentBlockedTest │ SUCCESS │ │ ├───────────────────────┼─────────┼─────────┤ │ Timeout │ SUCCESS │ │ ╘═══════════════════════╧═════════╧═════════╛

Registering the route statistics Registering the global statistics Finished all episode. Goodbye! KQKCTGIGT5P5468MXTXP63Z0P59KSNT5BZHAB92LYQXD0S3ENSEO3T90LX8G70TB PYTHON_RETURN=0!!! Start Over!!! Bash script done.

yixiao1 commented 2 weeks ago

Carla keeps crashing when I run Roach RL expert. Can you share the process of training and evaluating Roach RL expert? Thank you.

Hi, sorry for the late reply. We don’t train a new Roach RL expert agent but directly run the model provided in this GitHub. We used this agent as our expert agent to collect data, so training and evaluating it is not the focus for our paper. Please refer to their page and follow the instructions to run Roach RL agent.