eugenevinitsky/bayesian_reasoning_traffic issues and pull requests

#72 - Iclr branch

Pull Request - State: open - Opened by kevin-thankyou-lin about 4 years ago

#72 - Iclr branch

Pull Request - State: open - Opened by kevin-thankyou-lin about 4 years ago

#71 - Iclr branch:

Pull Request - State: closed - Opened by kevin-thankyou-lin about 4 years ago

#71 - Iclr branch:

Pull Request - State: closed - Opened by kevin-thankyou-lin about 4 years ago

#70 - ICLR-BRANCH: L2 actually stops at 40m

Pull Request - State: closed - Opened by kevin-thankyou-lin about 4 years ago

#70 - ICLR-BRANCH: L2 actually stops at 40m

Pull Request - State: closed - Opened by kevin-thankyou-lin about 4 years ago

#69 - Implement multi-lane training

Pull Request - State: open - Opened by kevin-thankyou-lin over 4 years ago

#69 - Implement multi-lane training

Pull Request - State: open - Opened by kevin-thankyou-lin over 4 years ago

#68 - push all files to klin/training

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#68 - push all files to klin/training

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#67 - added tensorboard to imitation

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#67 - added tensorboard to imitation

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#66 - Merge in rule-based controller improvement

Pull Request - State: closed - Opened by NickL77 over 4 years ago

#66 - Merge in rule-based controller improvement

Pull Request - State: closed - Opened by NickL77 over 4 years ago

#65 - Updated inference files

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#65 - Updated inference files

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#64 - Ev/inference testing

Pull Request - State: closed - Opened by eugenevinitsky over 4 years ago

#64 - Ev/inference testing

Pull Request - State: closed - Opened by eugenevinitsky over 4 years ago

#63 - Update ipynb for correct mean/sd

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#63 - Update ipynb for correct mean/sd

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#62 - inference using imitation network ready - network is self is crappy - training better one

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#62 - inference using imitation network ready - network is self is crappy - training better one

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#61 - Ev training branch to klin/training

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#61 - Ev training branch to klin/training

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#60 - Updated bayesian 0 to fully general network i.e random car starting p…

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#60 - Updated bayesian 0 to fully general network i.e random car starting p…

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#59 - ground truth pedestrian visibility state vector complete - still need…

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#59 - ground truth pedestrian visibility state vector complete - still need…

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#58 - Train DQN on ground truth to check for state space bug

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#58 - Train DQN on ground truth to check for state space bug

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#57 - remove imports and flags

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#57 - remove imports and flags

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#56 - Push for Eugene

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#56 - Push for Eugene

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#55 - Code cleanup env - NO NEED TO PULL YET

Pull Request - State: open - Opened by kevin-thankyou-lin over 4 years ago

#55 - Code cleanup env - NO NEED TO PULL YET

Pull Request - State: open - Opened by kevin-thankyou-lin over 4 years ago

#54 - Code cleanup

Pull Request - State: closed - Opened by eugenevinitsky over 4 years ago

#54 - Code cleanup

Pull Request - State: closed - Opened by eugenevinitsky over 4 years ago

#53 - Updated imitation and trained DQN

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#53 - Updated imitation and trained DQN

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#52 - Updated Bay0 for DQN

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#52 - Updated Bay0 for DQN

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#51 - update bay0nogrid

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#51 - update bay0nogrid

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#50 - DQN ready to train (updated reward section so obs and reward keys match) - Imitation learning has sample_weight bug

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#50 - DQN ready to train (updated reward section so obs and reward keys match) - Imitation learning has sample_weight bug

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#49 - Imitation learning - weight near intersection data - to debug 'sample weight' param

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#49 - Imitation learning - weight near intersection data - to debug 'sample weight' param

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#48 - Add option of entropy penalty (modify file from ray) + tune on penalty vals

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#48 - Add option of entropy penalty (modify file from ray) + tune on penalty vals

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#47 - Covariance matrix, entropy coeff to -0.01, continue sim if sumo car crashes

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#47 - Covariance matrix, entropy coeff to -0.01, continue sim if sumo car crashes

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#46 - Prediction on Bayesian 1

Pull Request - State: closed - Opened by NickL77 over 4 years ago

#46 - Prediction on Bayesian 1

Pull Request - State: closed - Opened by NickL77 over 4 years ago

#45 - Refactor inference notebook code

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#45 - Refactor inference notebook code

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#44 - push inference files to test

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#44 - push inference files to test

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#43 - shaped_97 has 800 iters and has good intersection approach behaviour,…

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#43 - shaped_97 has 800 iters and has good intersection approach behaviour,…

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#42 - Shape reward again

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#42 - Shape reward again

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#41 - ultimate reward shaping 200 250 300

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago - 1 comment

#41 - ultimate reward shaping 200 250 300

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago - 1 comment

#40 - reward shaping 2: distribute 'checkpoints' on the incoming edge

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#40 - reward shaping 2: distribute 'checkpoints' on the incoming edge

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#39 - Push the fix to github, round 2

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#39 - Push the fix to github, round 2

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#38 - properly fix the teleport bug

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#37 - fix teleportation-caused bug

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#37 - fix teleportation-caused bug

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#36 - reward shaping: reward getting 48-50m i,e 0-2m from intersection one time to encourage getting close earlier

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#36 - reward shaping: reward getting 48-50m i,e 0-2m from intersection one time to encourage getting close earlier

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#35 - increase veh collisions to -1500 (same as ped)

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#35 - increase veh collisions to -1500 (same as ped)

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#34 - adjusted ped wall and rewards

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#34 - adjusted ped wall and rewards

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#33 - Count pedestrian at a location only when its far enough along the corresponding walkway

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#32 - fixed bug in visible objects file 2

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#32 - fixed bug in visible objects file 2

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#31 - fixed bug in visible objects file

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#31 - fixed bug in visible objects file

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#30 - 4 ped locations ready

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#30 - 4 ped locations ready

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#29 - 4 ped location state

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#28 - Merge local changes

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#28 - Merge local changes

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#27 - prepare code for 1 ped scheme: randomize top car's destination + give…

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#27 - prepare code for 1 ped scheme: randomize top car's destination + give…

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#26 - increase horizon to 1200ms so rl vehicle can cross intersection after pedestrian wave + simple optimizer = False

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#26 - increase horizon to 1200ms so rl vehicle can cross intersection after pedestrian wave + simple optimizer = False

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#25 - prepare bay 0 no grid for truly random peds this time

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#25 - prepare bay 0 no grid for truly random peds this time

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#24 - updated bayesian 0 no grid env so that rl doesn't just learn to deter…

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#24 - updated bayesian 0 no grid env so that rl doesn't just learn to deter…

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#23 - prepare bayesian 0 no grid for training

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#23 - prepare bayesian 0 no grid for training

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#22 - Bayesian 0 no grid

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#22 - Bayesian 0 no grid

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

#21 - Change penalty for bay env 1 hard braking to actual hard braking i.e.…

Pull Request - State: closed - Opened by kevin-thankyou-lin over 4 years ago

GitHub / eugenevinitsky/bayesian_reasoning_traffic issues and pull requests