Closed thoithoi58 closed 1 year ago
@thoithoi58, after you delete outputs/checkpoint.txt
, run train_rl.sh. Make sure you let enough training take place that something syncs with wandb. Copy the run_id from outputs/checkpoint.txt
to rl_birdview_agent.py.
Here's my code: https://github.com/neilsambhu/carla-roach/blob/NeilBranch0/timeline/README3.md Go to timestamps (1) 6/19/2022 5:44 PM and (2) 6/20/2022 11:50 AM
Here is my rl_birdview_agent.py: https://github.com/neilsambhu/carla-roach/blob/NeilBranch0/agents/rl_birdview/rl_birdview_agent.py#L107
As of writing this message, I'm still waiting for my initial training to finish to see if my code to read in the run_id from the outputs/checkpoint.txt
works.
Hi, I'm running
run/train_rl.sh
and keep receiving this errorI've browsed the issues page and found the same error from other person here, and the solution is delete the
outputs/checkpoint.txt
. But to me it was no help