4 bernhard dev - #570
Merged
Merged
Conversation
env.goal_source=route \
env.max_goal_spacing=60 \
…ed on the date and I do not need to touch the script anymore.,
Victorbares
reviewed
Aug 19, 2026
…ldens Tests still called check_stop_line_crossing after its rename to check_agent_on_stop_line. Smoke goldens regenerated in the pinned QEMU image to reflect the intentional red light logic and observation changes on this branch.
eugenevinitsky
approved these changes
Aug 20, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Make reward_log_sampling, and reset_accel_on_stop optional and turned off by default. (Makes current training run unstable).
Make increasing ego box configurable for eval eval_perceived_size_margin_m
I did not change the default.
Add option to add layer norm to the heads as well. Turned off by default
Fix bug where the input items are clipped if height is difference is more than 4 meters. This removed nearby cars on steep slopes in town03. I removed this filter because the ego get x,y,z as input so can judge by itself if a car is not relevant.
Fix some visu bugs and compiler warnings
Only keep the lastest model checkpoint and delete the older once to save disk space.
Fix red light bug where the model got red light infractions when inside intersections. Now only infractions are incurred when the model enters an intersection.
Validated with a 1T run, which achieves a new record of 10k km/infraction in self-play
nuPlan numbers did not improve/reduced a bit. (I think that is fine for now since nuPlan evals are still full of bugs and we want to get the Carla evals / learning clean and good first before we fix generalization.)
https://wandb.ai/emerge_/nightly-multi-long/runs/k_scaled_0010_1000