Gsp dev #149
Closed
Gsp dev #149
+3,900,272
−110,282
Conversation
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
* Add human ll loss with random actions [placeholder]. * Change clip coef * Typo * Integrate the actual human data into pipeline. * Back at 194 K FPS. Save data to disk, only load small batch to gpu when needed. * Ignore human data. * Revert small unrelated change. * Remove print statement * Return the continuous actions for sanity checks. * Log the weighted loss objectives instead. * Always compute human log prob for systematic analysis. * Bug fix: do not use human entropy. * Small fixes * Experimental settings. * Simplify expert data generation logic. * Small improvement * Store amount of human data used in checkpoints (each seq is 3 seconds of human data). * Cleanup/
* WIP guidance reward * remove irrelevant const * heading fix * minor refactor * fix normalize function defn * removed master weight * guidance reward sweeps
…#112) * start on obs space cfg * rmv magic numbers * Pass ini -> drive env. * Working on fuse linear * Integrate 700k SPS kernel. * Dirty * Profile fix * linear max kernel * Add linearmax kernels * LinearMax with layernorm: 600/700K SPS, score of 0.85 in < 5 min. * Use 1000 maps * Best run gets score of .95 in 25 minutes. * Advantage weighting exps * Sweep setup. * Increase steps. * Remove exploration logging. * Delete Triton kernel. * Best score is 0.98 with 600K FPS. 800 K FPS with 128 road points. * Better hparams * Replace default learning rate and ent_coef. * Minor * start on obs space cfg * rmv magic numbers * Pass ini -> drive env. * Working on fuse linear * Integrate 700k SPS kernel. * Dirty * Profile fix * linear max kernel * Add linearmax kernels * LinearMax with layernorm: 600/700K SPS, score of 0.85 in < 5 min. * Best run gets score of .95 in 25 minutes. * Advantage weighting exps * Sweep setup. * Increase steps. * Remove exploration logging. * Delete Triton kernel. * Best score is 0.98 with 600K FPS. 800 K FPS with 128 road points. * Better hparams * Clean up. * Minor * Update profiling script. * Adapt drive c network. * Temp settings * minor * Set max agents to 32 * Add linear -> relu -> max kernel. * Score is .98 after 70 minutes. * Minor --------- Co-authored-by: Joseph Suarez <jsuarez@puffer.ai>
* Add model binary file and make demo() work. * Add docs. * Add docs. --------- Co-authored-by: Daphne <daphn3cor@gmail.com>
* Added Z Coords, Polygonized Junction Area to handle point in polygon query * Added Snapped Polylines to better polygonizing * Fixed Extra Road Lines bug, better polygonization with debugging * Fixed initial heading angle for Trajectory sampling * Maps Before manual filtering * Carla Before Manual with z coordinates * NaN fixes * Minor * Carla Maps Cleaned 6/8 * Add pyxodr submodule at external/pyxodr * added external submodule support, fixed num_tries for valid init position of agents, added arg parsing, cleaned up code * Removed unstable process_roads_script, use carla_py123d * add avg_speed as arg * Remove old Carla Maps * Update README.md Co-authored-by: greptile-apps[bot] <165735046+greptile-apps[bot]@users.noreply.github.com> * Remove old jsons and testing files * Remove redundant instructions from README * indentation changes * Minor editorial edits. --------- Co-authored-by: Daphne Cornelisse <cor.daphne@gmail.com> Co-authored-by: greptile-apps[bot] <165735046+greptile-apps[bot]@users.noreply.github.com> Co-authored-by: Daphne <daphn3cor@gmail.com>
* collision fix * lowered to 250 for actual theoretical guarantee
* initial commit * remove function * restore weights, add trailing newline, fix ini * update to unify logic and refactor to function * fix reverse computation * precommit fix --------- Co-authored-by: Kevin <kj2676@nyu.edu>
* Documentation first pass * incorporate previous docs * style updates and corrections * initial commit * Format drive helpers
* Adding collision behavior back * Removing uneccesary snippets * Rebased * precommit fixes * Pre-Commit fixes --------- Co-authored-by: Daphne Cornelisse <cor.daphne@gmail.com>
* Added valid 2d carla towns, some code cleanup * Add carla 3D maps with objects
…pling in longer episodes (#186) * Make road lines and lanes visible in map. * Simplify goal resample algorithm: Pick best road lane point in road graph. * Delete redundant code. * Make the target distance to the new goal configurable. * Generalize metrics to work for longer episodes with resampling. Also delete a bunch of unused graph topology code. * Minor * Apply precommit. * Fix in visualizer. * fix metrics * WIP * Add goal behavior flag. * Add fallback for goal resampling and cleanup. * Make goal radius more visible. * Minor * Make grid appear in the background. * Minor. * Merge * Fix bug in logging num goals reached and sampled. * Add goal taret * Use classic dynamics model. * Fix descrepancies between demo() and eval_gif(). * Small bug fix. * Reward shaping * Termination mode must be 0 for Carla maps. * Add all args from ini to demo() env. * Clean up visualization code. * Clean up metrics and vis. * Fix metrics. * Add diversity to agent view. * Add better fallback. * Reserve red for cars that are in collision. * Keep track of current goals. * Carla testing simple/ * Use classic dynamics by default. * Fix small bug in goal logging (respawn). * Always draw agent obs when resampling goals. * Increase render videos timeout (carla maps take longer). * Minor vis changes. * Minor vis changes. * Rmv displacement error for now and add goal speed target. * Add optional goal speed. * Incorporate suggestions. * Revert settings. * Revert settings. * Revert settings. * Fixes * Add docs * Minor * Make grid appear in background. * Edits. * Typo. * Minor visual adaptations. --------- Co-authored-by: Daphne <daphn3cor@gmail.com> Co-authored-by: julianh65 <jhunt17159@gmail.com>
* Corrections of the resample code in drive.py: - the will_resample=1 followed by if will_resample looked weird to me (probably legacy code ?) - When we resample we should update the values of self.agent_offsets map dirs and num envs. The fact that we didn't update them isn't an issue because right now they are not accessed anywhere in the code, but then we should either remove these attributes of the Drive Class or either make ensure they contain the right values if someone wants to use them later. * Minor * Fix merge conflicts. --------- Co-authored-by: Daphne Cornelisse <cor.daphne@gmail.com>
…del (#206) * Fix human control with joint action space & classic model: Was still assuming multi-discrete. * Enable human control with jerks dynamics model. * Color actions yellow when controlling. * Slightly easier control problem? * Add tiny jerk penalty: Results in smooth behavior. * Pre-commit * Minor edits. * Revert ini changes. --------- Co-authored-by: Daphne <daphn3cor@gmail.com>
* Added WOSAC results on the 10k validation dataset * Code to evaluate SMART + associated doc * Edits. * Add link to docs. --------- Co-authored-by: Wael Boumediene Doulazmi <wbd2016@gl001.hpc.nyu.edu> Co-authored-by: Daphne Cornelisse <cor.daphne@gmail.com>
* Good behavior with trained policy - resampling. * Hardcode to 10 max agents for web version. * Browser demo v1. * More descriptive docs. * Release post edits. * Docs improvements. * Run precommit. * Better policy. * Revert .ini changes, except one. * Delete drive.dSYM/Contents/Info.plist * Delete pufferlib/resources/drive/puffer_drive_gljhhrl6.bin --------- Co-authored-by: Daphne <daphn3cor@gmail.com>
Co-authored-by: Daphne <daphn3cor@gmail.com>
Co-authored-by: Daphne <daphn3cor@gmail.com>
Move from mkdocs to mdbooks. Code heavily claude assisted.
Co-authored-by: Daphne <daphn3cor@gmail.com>
* fix minor errors * try to fix things for dark mode * Fixing dark/light mode error --------- Co-authored-by: Eugene Vinitsky <eugene@percepta.ai> Co-authored-by: Aditya Gupta <adigupta2602@gmail.com>
* Ensure the arrows are on the ground. * Date fix. * Update data docs with mixed dataset. * Increase map range. * Remove ini file for web demo. * WIP * Edit release post. * Minor docs fixes. * Writing changes. * Fix metrics. * Delete outdated readme files. * Minor. * Improve docs. * Fix institutions. * Update training info. * Reset configs to womd defaults. * Update score metrics. * Update docs * Update policy for demo. * Update demo files (new cpt) * Minor. * Add video. * Minor. * Keep defaults. --------- Co-authored-by: Daphne <daphn3cor@gmail.com>
Authored by vinitsky.eugene@gmail.com
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
No description provided.