From 338437616e57715b9c2e649b9b68eb7a53fd6983 Mon Sep 17 00:00:00 2001 From: Kate Tolstaya Date: Wed, 24 Jul 2019 22:12:45 +0100 Subject: [PATCH] cleaning --- README.md | 11 +- cfg/cloning.cfg | 11 +- cfg/dagger.cfg | 2 +- cfg/old/n.cfg | 834 ------------------ cfg/old/n_baseline.cfg | 434 --------- cfg/old/rad.cfg | 179 ---- cfg/old/rad2.cfg | 832 ----------------- cfg/old/rad_baseline.cfg | 434 --------- cfg/old/vel.cfg | 201 ----- cfg/old/vel2.cfg | 832 ----------------- cfg/old/vel_baseline.cfg | 434 --------- learner/gnn_cloning.py | 2 +- learner/gnn_dagger.py | 2 +- models/actor_FlockingRelative-v0_dagger_k3 | Bin 0 -> 7990 bytes .../figure_code/test_3pane.py | 0 .../figure_code/test_one_traj.py | 0 .../figure_code/test_one_traj2.py | 0 test_model.py | 16 +- 18 files changed, 22 insertions(+), 4202 deletions(-) delete mode 100644 cfg/old/n.cfg delete mode 100644 cfg/old/n_baseline.cfg delete mode 100644 cfg/old/rad.cfg delete mode 100644 cfg/old/rad2.cfg delete mode 100644 cfg/old/rad_baseline.cfg delete mode 100644 cfg/old/vel.cfg delete mode 100644 cfg/old/vel2.cfg delete mode 100644 cfg/old/vel_baseline.cfg create mode 100644 models/actor_FlockingRelative-v0_dagger_k3 rename test_3pane.py => results/figure_code/test_3pane.py (100%) rename test_one_traj.py => results/figure_code/test_one_traj.py (100%) rename test_one_traj2.py => results/figure_code/test_one_traj2.py (100%) diff --git a/README.md b/README.md index 70dff1a..430694f 100644 --- a/README.md +++ b/README.md @@ -8,14 +8,11 @@ ## Available algorithms: - Behavior Cloning as described in [ArXiv](https://arxiv.org/abs/1903.10527) `python3 train.py cfg/cloning.cfg` - DAGGER imitation learning `python3 train.py cfg/dagger.cfg` -- Deep Deterministic Policy Gradients (TODO) + +## To test: +- `python3 test_model.py cfg/dagger.cfg` ## Other code: - `python3 flocking_gym_test.py` provides test code for the Gym Flock environments -## To Do -- Plot centralized and decentralized baselines -- Obstacle avoidance -- Follow 1 or 2 leaders -- Play with control costs -- DDPG + diff --git a/cfg/cloning.cfg b/cfg/cloning.cfg index 99d2fd1..a1d71fd 100644 --- a/cfg/cloning.cfg +++ b/cfg/cloning.cfg @@ -9,6 +9,11 @@ updates_per_step = 200 seed = 11 actor_lr = 5e-5 +n_train_episodes = 400 +beta_coeff = 0.993 +test_interval = 40 +n_test_episodes = 20 + # architecture parameters k = 2 hidden_size = 32 @@ -22,7 +27,9 @@ comm_radius = 1.0 n_agents = 100 n_actions = 2 n_states = 6 -debug = False +debug = True header = reward +dt = 0.01 -[test] \ No newline at end of file +[test] +fname = cloning_k3 \ No newline at end of file diff --git a/cfg/dagger.cfg b/cfg/dagger.cfg index aff8cf7..60c8b95 100644 --- a/cfg/dagger.cfg +++ b/cfg/dagger.cfg @@ -33,4 +33,4 @@ dt = 0.01 [test] -fname = k3 \ No newline at end of file +fname = dagger_k3 \ No newline at end of file diff --git a/cfg/old/n.cfg b/cfg/old/n.cfg deleted file mode 100644 index 744470d..0000000 --- a/cfg/old/n.cfg +++ /dev/null @@ -1,834 +0,0 @@ -[DEFAULT] - -alg = dagger - -# learning parameters -batch_size = 20 -buffer_size = 10000 -updates_per_step = 200 -seed = 11 -actor_lr = 5e-5 - -n_train_episodes = 800 -beta_coeff = 0.993 -test_interval = 40 -n_test_episodes = 20 - -# architecture parameters -k = 2 -hidden_size = 32 -gamma = 0.99 -tau = 0.5 - -# env parameters -env = FlockingRelative-v0 -v_max = 3.0 -comm_radius = 1.0 -n_agents = 100 -n_actions = 2 -n_states = 6 -debug = False - - -header = k, seed, n_agents, reward - -[1, 0, 20] -k = 1 -seed = 0 -n_agents = 20 - -[1, 0, 40] -k = 1 -seed = 0 -n_agents = 40 - -[1, 0, 80] -k = 1 -seed = 0 -n_agents = 80 - -[1, 0, 100] -k = 1 -seed = 0 -n_agents = 100 - -[1, 1, 20] -k = 1 -seed = 1 -n_agents = 20 - -[1, 1, 40] -k = 1 -seed = 1 -n_agents = 40 - -[1, 1, 80] -k = 1 -seed = 1 -n_agents = 80 - -[1, 1, 100] -k = 1 -seed = 1 -n_agents = 100 - -[1, 2, 20] -k = 1 -seed = 2 -n_agents = 20 - -[1, 2, 40] -k = 1 -seed = 2 -n_agents = 40 - -[1, 2, 80] -k = 1 -seed = 2 -n_agents = 80 - -[1, 2, 100] -k = 1 -seed = 2 -n_agents = 100 - -[1, 3, 20] -k = 1 -seed = 3 -n_agents = 20 - -[1, 3, 40] -k = 1 -seed = 3 -n_agents = 40 - -[1, 3, 80] -k = 1 -seed = 3 -n_agents = 80 - -[1, 3, 100] -k = 1 -seed = 3 -n_agents = 100 - -[1, 4, 20] -k = 1 -seed = 4 -n_agents = 20 - -[1, 4, 40] -k = 1 -seed = 4 -n_agents = 40 - -[1, 4, 80] -k = 1 -seed = 4 -n_agents = 80 - -[1, 4, 100] -k = 1 -seed = 4 -n_agents = 100 - -[1, 5, 20] -k = 1 -seed = 5 -n_agents = 20 - -[1, 5, 40] -k = 1 -seed = 5 -n_agents = 40 - -[1, 5, 80] -k = 1 -seed = 5 -n_agents = 80 - -[1, 5, 100] -k = 1 -seed = 5 -n_agents = 100 - -[1, 6, 20] -k = 1 -seed = 6 -n_agents = 20 - -[1, 6, 40] -k = 1 -seed = 6 -n_agents = 40 - -[1, 6, 80] -k = 1 -seed = 6 -n_agents = 80 - -[1, 6, 100] -k = 1 -seed = 6 -n_agents = 100 - -[1, 7, 20] -k = 1 -seed = 7 -n_agents = 20 - -[1, 7, 40] -k = 1 -seed = 7 -n_agents = 40 - -[1, 7, 80] -k = 1 -seed = 7 -n_agents = 80 - -[1, 7, 100] -k = 1 -seed = 7 -n_agents = 100 - -[1, 8, 20] -k = 1 -seed = 8 -n_agents = 20 - -[1, 8, 40] -k = 1 -seed = 8 -n_agents = 40 - -[1, 8, 80] -k = 1 -seed = 8 -n_agents = 80 - -[1, 8, 100] -k = 1 -seed = 8 -n_agents = 100 - -[1, 9, 20] -k = 1 -seed = 9 -n_agents = 20 - -[1, 9, 40] -k = 1 -seed = 9 -n_agents = 40 - -[1, 9, 80] -k = 1 -seed = 9 -n_agents = 80 - -[1, 9, 100] -k = 1 -seed = 9 -n_agents = 100 - -[2, 0, 20] -k = 2 -seed = 0 -n_agents = 20 - -[2, 0, 40] -k = 2 -seed = 0 -n_agents = 40 - -[2, 0, 80] -k = 2 -seed = 0 -n_agents = 80 - -[2, 0, 100] -k = 2 -seed = 0 -n_agents = 100 - -[2, 1, 20] -k = 2 -seed = 1 -n_agents = 20 - -[2, 1, 40] -k = 2 -seed = 1 -n_agents = 40 - -[2, 1, 80] -k = 2 -seed = 1 -n_agents = 80 - -[2, 1, 100] -k = 2 -seed = 1 -n_agents = 100 - -[2, 2, 20] -k = 2 -seed = 2 -n_agents = 20 - -[2, 2, 40] -k = 2 -seed = 2 -n_agents = 40 - -[2, 2, 80] -k = 2 -seed = 2 -n_agents = 80 - -[2, 2, 100] -k = 2 -seed = 2 -n_agents = 100 - -[2, 3, 20] -k = 2 -seed = 3 -n_agents = 20 - -[2, 3, 40] -k = 2 -seed = 3 -n_agents = 40 - -[2, 3, 80] -k = 2 -seed = 3 -n_agents = 80 - -[2, 3, 100] -k = 2 -seed = 3 -n_agents = 100 - -[2, 4, 20] -k = 2 -seed = 4 -n_agents = 20 - -[2, 4, 40] -k = 2 -seed = 4 -n_agents = 40 - -[2, 4, 80] -k = 2 -seed = 4 -n_agents = 80 - -[2, 4, 100] -k = 2 -seed = 4 -n_agents = 100 - -[2, 5, 20] -k = 2 -seed = 5 -n_agents = 20 - -[2, 5, 40] -k = 2 -seed = 5 -n_agents = 40 - -[2, 5, 80] -k = 2 -seed = 5 -n_agents = 80 - -[2, 5, 100] -k = 2 -seed = 5 -n_agents = 100 - -[2, 6, 20] -k = 2 -seed = 6 -n_agents = 20 - -[2, 6, 40] -k = 2 -seed = 6 -n_agents = 40 - -[2, 6, 80] -k = 2 -seed = 6 -n_agents = 80 - -[2, 6, 100] -k = 2 -seed = 6 -n_agents = 100 - -[2, 7, 20] -k = 2 -seed = 7 -n_agents = 20 - -[2, 7, 40] -k = 2 -seed = 7 -n_agents = 40 - -[2, 7, 80] -k = 2 -seed = 7 -n_agents = 80 - -[2, 7, 100] -k = 2 -seed = 7 -n_agents = 100 - -[2, 8, 20] -k = 2 -seed = 8 -n_agents = 20 - -[2, 8, 40] -k = 2 -seed = 8 -n_agents = 40 - -[2, 8, 80] -k = 2 -seed = 8 -n_agents = 80 - -[2, 8, 100] -k = 2 -seed = 8 -n_agents = 100 - -[2, 9, 20] -k = 2 -seed = 9 -n_agents = 20 - -[2, 9, 40] -k = 2 -seed = 9 -n_agents = 40 - -[2, 9, 80] -k = 2 -seed = 9 -n_agents = 80 - -[2, 9, 100] -k = 2 -seed = 9 -n_agents = 100 - -[3, 0, 20] -k = 3 -seed = 0 -n_agents = 20 - -[3, 0, 40] -k = 3 -seed = 0 -n_agents = 40 - -[3, 0, 80] -k = 3 -seed = 0 -n_agents = 80 - -[3, 0, 100] -k = 3 -seed = 0 -n_agents = 100 - -[3, 1, 20] -k = 3 -seed = 1 -n_agents = 20 - -[3, 1, 40] -k = 3 -seed = 1 -n_agents = 40 - -[3, 1, 80] -k = 3 -seed = 1 -n_agents = 80 - -[3, 1, 100] -k = 3 -seed = 1 -n_agents = 100 - -[3, 2, 20] -k = 3 -seed = 2 -n_agents = 20 - -[3, 2, 40] -k = 3 -seed = 2 -n_agents = 40 - -[3, 2, 80] -k = 3 -seed = 2 -n_agents = 80 - -[3, 2, 100] -k = 3 -seed = 2 -n_agents = 100 - -[3, 3, 20] -k = 3 -seed = 3 -n_agents = 20 - -[3, 3, 40] -k = 3 -seed = 3 -n_agents = 40 - -[3, 3, 80] -k = 3 -seed = 3 -n_agents = 80 - -[3, 3, 100] -k = 3 -seed = 3 -n_agents = 100 - -[3, 4, 20] -k = 3 -seed = 4 -n_agents = 20 - -[3, 4, 40] -k = 3 -seed = 4 -n_agents = 40 - -[3, 4, 80] -k = 3 -seed = 4 -n_agents = 80 - -[3, 4, 100] -k = 3 -seed = 4 -n_agents = 100 - -[3, 5, 20] -k = 3 -seed = 5 -n_agents = 20 - -[3, 5, 40] -k = 3 -seed = 5 -n_agents = 40 - -[3, 5, 80] -k = 3 -seed = 5 -n_agents = 80 - -[3, 5, 100] -k = 3 -seed = 5 -n_agents = 100 - -[3, 6, 20] -k = 3 -seed = 6 -n_agents = 20 - -[3, 6, 40] -k = 3 -seed = 6 -n_agents = 40 - -[3, 6, 80] -k = 3 -seed = 6 -n_agents = 80 - -[3, 6, 100] -k = 3 -seed = 6 -n_agents = 100 - -[3, 7, 20] -k = 3 -seed = 7 -n_agents = 20 - -[3, 7, 40] -k = 3 -seed = 7 -n_agents = 40 - -[3, 7, 80] -k = 3 -seed = 7 -n_agents = 80 - -[3, 7, 100] -k = 3 -seed = 7 -n_agents = 100 - -[3, 8, 20] -k = 3 -seed = 8 -n_agents = 20 - -[3, 8, 40] -k = 3 -seed = 8 -n_agents = 40 - -[3, 8, 80] -k = 3 -seed = 8 -n_agents = 80 - -[3, 8, 100] -k = 3 -seed = 8 -n_agents = 100 - -[3, 9, 20] -k = 3 -seed = 9 -n_agents = 20 - -[3, 9, 40] -k = 3 -seed = 9 -n_agents = 40 - -[3, 9, 80] -k = 3 -seed = 9 -n_agents = 80 - -[3, 9, 100] -k = 3 -seed = 9 -n_agents = 100 - -[4, 0, 20] -k = 4 -seed = 0 -n_agents = 20 - -[4, 0, 40] -k = 4 -seed = 0 -n_agents = 40 - -[4, 0, 80] -k = 4 -seed = 0 -n_agents = 80 - -[4, 0, 100] -k = 4 -seed = 0 -n_agents = 100 - -[4, 1, 20] -k = 4 -seed = 1 -n_agents = 20 - -[4, 1, 40] -k = 4 -seed = 1 -n_agents = 40 - -[4, 1, 80] -k = 4 -seed = 1 -n_agents = 80 - -[4, 1, 100] -k = 4 -seed = 1 -n_agents = 100 - -[4, 2, 20] -k = 4 -seed = 2 -n_agents = 20 - -[4, 2, 40] -k = 4 -seed = 2 -n_agents = 40 - -[4, 2, 80] -k = 4 -seed = 2 -n_agents = 80 - -[4, 2, 100] -k = 4 -seed = 2 -n_agents = 100 - -[4, 3, 20] -k = 4 -seed = 3 -n_agents = 20 - -[4, 3, 40] -k = 4 -seed = 3 -n_agents = 40 - -[4, 3, 80] -k = 4 -seed = 3 -n_agents = 80 - -[4, 3, 100] -k = 4 -seed = 3 -n_agents = 100 - -[4, 4, 20] -k = 4 -seed = 4 -n_agents = 20 - -[4, 4, 40] -k = 4 -seed = 4 -n_agents = 40 - -[4, 4, 80] -k = 4 -seed = 4 -n_agents = 80 - -[4, 4, 100] -k = 4 -seed = 4 -n_agents = 100 - -[4, 5, 20] -k = 4 -seed = 5 -n_agents = 20 - -[4, 5, 40] -k = 4 -seed = 5 -n_agents = 40 - -[4, 5, 80] -k = 4 -seed = 5 -n_agents = 80 - -[4, 5, 100] -k = 4 -seed = 5 -n_agents = 100 - -[4, 6, 20] -k = 4 -seed = 6 -n_agents = 20 - -[4, 6, 40] -k = 4 -seed = 6 -n_agents = 40 - -[4, 6, 80] -k = 4 -seed = 6 -n_agents = 80 - -[4, 6, 100] -k = 4 -seed = 6 -n_agents = 100 - -[4, 7, 20] -k = 4 -seed = 7 -n_agents = 20 - -[4, 7, 40] -k = 4 -seed = 7 -n_agents = 40 - -[4, 7, 80] -k = 4 -seed = 7 -n_agents = 80 - -[4, 7, 100] -k = 4 -seed = 7 -n_agents = 100 - -[4, 8, 20] -k = 4 -seed = 8 -n_agents = 20 - -[4, 8, 40] -k = 4 -seed = 8 -n_agents = 40 - -[4, 8, 80] -k = 4 -seed = 8 -n_agents = 80 - -[4, 8, 100] -k = 4 -seed = 8 -n_agents = 100 - -[4, 9, 20] -k = 4 -seed = 9 -n_agents = 20 - -[4, 9, 40] -k = 4 -seed = 9 -n_agents = 40 - -[4, 9, 80] -k = 4 -seed = 9 -n_agents = 80 - -[4, 9, 100] -k = 4 -seed = 9 -n_agents = 100 - diff --git a/cfg/old/n_baseline.cfg b/cfg/old/n_baseline.cfg deleted file mode 100644 index 30bb616..0000000 --- a/cfg/old/n_baseline.cfg +++ /dev/null @@ -1,434 +0,0 @@ -[DEFAULT] - -alg = baseline - -# learning parameters -batch_size = 20 -buffer_size = 10000 -updates_per_step = 200 -seed = 11 -actor_lr = 5e-5 - -n_train_episodes = 800 -beta_coeff = 0.993 -test_interval = 40 -n_test_episodes = 20 - -# architecture parameters -k = 2 -hidden_size = 32 -gamma = 0.99 -tau = 0.5 - -# env parameters -env = FlockingRelative-v0 -v_max = 3.0 -comm_radius = 1.0 -n_agents = 100 -n_actions = 2 -n_states = 6 -debug = False - - -header = seed, centralized, n_agents, reward - -[0, True, 20] -seed = 0 -centralized = True -n_agents = 20 - -[0, True, 40] -seed = 0 -centralized = True -n_agents = 40 - -[0, True, 80] -seed = 0 -centralized = True -n_agents = 80 - -[0, True, 100] -seed = 0 -centralized = True -n_agents = 100 - -[0, False, 20] -seed = 0 -centralized = False -n_agents = 20 - -[0, False, 40] -seed = 0 -centralized = False -n_agents = 40 - -[0, False, 80] -seed = 0 -centralized = False -n_agents = 80 - -[0, False, 100] -seed = 0 -centralized = False -n_agents = 100 - -[1, True, 20] -seed = 1 -centralized = True -n_agents = 20 - -[1, True, 40] -seed = 1 -centralized = True -n_agents = 40 - -[1, True, 80] -seed = 1 -centralized = True -n_agents = 80 - -[1, True, 100] -seed = 1 -centralized = True -n_agents = 100 - -[1, False, 20] -seed = 1 -centralized = False -n_agents = 20 - -[1, False, 40] -seed = 1 -centralized = False -n_agents = 40 - -[1, False, 80] -seed = 1 -centralized = False -n_agents = 80 - -[1, False, 100] -seed = 1 -centralized = False -n_agents = 100 - -[2, True, 20] -seed = 2 -centralized = True -n_agents = 20 - -[2, True, 40] -seed = 2 -centralized = True -n_agents = 40 - -[2, True, 80] -seed = 2 -centralized = True -n_agents = 80 - -[2, True, 100] -seed = 2 -centralized = True -n_agents = 100 - -[2, False, 20] -seed = 2 -centralized = False -n_agents = 20 - -[2, False, 40] -seed = 2 -centralized = False -n_agents = 40 - -[2, False, 80] -seed = 2 -centralized = False -n_agents = 80 - -[2, False, 100] -seed = 2 -centralized = False -n_agents = 100 - -[3, True, 20] -seed = 3 -centralized = True -n_agents = 20 - -[3, True, 40] -seed = 3 -centralized = True -n_agents = 40 - -[3, True, 80] -seed = 3 -centralized = True -n_agents = 80 - -[3, True, 100] -seed = 3 -centralized = True -n_agents = 100 - -[3, False, 20] -seed = 3 -centralized = False -n_agents = 20 - -[3, False, 40] -seed = 3 -centralized = False -n_agents = 40 - -[3, False, 80] -seed = 3 -centralized = False -n_agents = 80 - -[3, False, 100] -seed = 3 -centralized = False -n_agents = 100 - -[4, True, 20] -seed = 4 -centralized = True -n_agents = 20 - -[4, True, 40] -seed = 4 -centralized = True -n_agents = 40 - -[4, True, 80] -seed = 4 -centralized = True -n_agents = 80 - -[4, True, 100] -seed = 4 -centralized = True -n_agents = 100 - -[4, False, 20] -seed = 4 -centralized = False -n_agents = 20 - -[4, False, 40] -seed = 4 -centralized = False -n_agents = 40 - -[4, False, 80] -seed = 4 -centralized = False -n_agents = 80 - -[4, False, 100] -seed = 4 -centralized = False -n_agents = 100 - -[5, True, 20] -seed = 5 -centralized = True -n_agents = 20 - -[5, True, 40] -seed = 5 -centralized = True -n_agents = 40 - -[5, True, 80] -seed = 5 -centralized = True -n_agents = 80 - -[5, True, 100] -seed = 5 -centralized = True -n_agents = 100 - -[5, False, 20] -seed = 5 -centralized = False -n_agents = 20 - -[5, False, 40] -seed = 5 -centralized = False -n_agents = 40 - -[5, False, 80] -seed = 5 -centralized = False -n_agents = 80 - -[5, False, 100] -seed = 5 -centralized = False -n_agents = 100 - -[6, True, 20] -seed = 6 -centralized = True -n_agents = 20 - -[6, True, 40] -seed = 6 -centralized = True -n_agents = 40 - -[6, True, 80] -seed = 6 -centralized = True -n_agents = 80 - -[6, True, 100] -seed = 6 -centralized = True -n_agents = 100 - -[6, False, 20] -seed = 6 -centralized = False -n_agents = 20 - -[6, False, 40] -seed = 6 -centralized = False -n_agents = 40 - -[6, False, 80] -seed = 6 -centralized = False -n_agents = 80 - -[6, False, 100] -seed = 6 -centralized = False -n_agents = 100 - -[7, True, 20] -seed = 7 -centralized = True -n_agents = 20 - -[7, True, 40] -seed = 7 -centralized = True -n_agents = 40 - -[7, True, 80] -seed = 7 -centralized = True -n_agents = 80 - -[7, True, 100] -seed = 7 -centralized = True -n_agents = 100 - -[7, False, 20] -seed = 7 -centralized = False -n_agents = 20 - -[7, False, 40] -seed = 7 -centralized = False -n_agents = 40 - -[7, False, 80] -seed = 7 -centralized = False -n_agents = 80 - -[7, False, 100] -seed = 7 -centralized = False -n_agents = 100 - -[8, True, 20] -seed = 8 -centralized = True -n_agents = 20 - -[8, True, 40] -seed = 8 -centralized = True -n_agents = 40 - -[8, True, 80] -seed = 8 -centralized = True -n_agents = 80 - -[8, True, 100] -seed = 8 -centralized = True -n_agents = 100 - -[8, False, 20] -seed = 8 -centralized = False -n_agents = 20 - -[8, False, 40] -seed = 8 -centralized = False -n_agents = 40 - -[8, False, 80] -seed = 8 -centralized = False -n_agents = 80 - -[8, False, 100] -seed = 8 -centralized = False -n_agents = 100 - -[9, True, 20] -seed = 9 -centralized = True -n_agents = 20 - -[9, True, 40] -seed = 9 -centralized = True -n_agents = 40 - -[9, True, 80] -seed = 9 -centralized = True -n_agents = 80 - -[9, True, 100] -seed = 9 -centralized = True -n_agents = 100 - -[9, False, 20] -seed = 9 -centralized = False -n_agents = 20 - -[9, False, 40] -seed = 9 -centralized = False -n_agents = 40 - -[9, False, 80] -seed = 9 -centralized = False -n_agents = 80 - -[9, False, 100] -seed = 9 -centralized = False -n_agents = 100 - diff --git a/cfg/old/rad.cfg b/cfg/old/rad.cfg deleted file mode 100644 index a132f76..0000000 --- a/cfg/old/rad.cfg +++ /dev/null @@ -1,179 +0,0 @@ -[DEFAULT] - -alg = dagger - -# learning parameters -batch_size = 20 -buffer_size = 10000 -updates_per_step = 200 -seed = 11 -actor_lr = 5e-5 - -# architecture parameters -k = 2 -hidden_size = 32 -gamma = 0.99 -tau = 0.5 - -# env parameters -env = FlockingRelative-v0 -v_max = 3.0 -comm_radius = 1.0 -n_agents = 100 -n_actions = 2 -n_states = 6 -debug = False -header = comm_radius, seed, k, reward - - -[3.0, 1, 1] -comm_radius = 3.0 -seed = 1 -k = 1 - -[2.0, 1, 1] -comm_radius = 2.0 -seed = 1 -k = 1 - -[1.0, 1, 1] -comm_radius = 1.0 -seed = 1 -k = 1 - -## - -[3.0, 2, 1] -comm_radius = 3.0 -seed = 1 -k = 2 - -[2.0, 2, 1] -comm_radius = 2.0 -seed = 1 -k = 2 - -[1.0, 2, 1] -comm_radius = 1.0 -seed = 1 -k = 2 - -## - -[3.0, 3, 1] -comm_radius = 3.0 -seed = 1 -k = 3 - -[2.0, 3, 1] -comm_radius = 2.0 -seed = 1 -k = 3 - -[1.0, 3, 1] -comm_radius = 1.0 -seed = 1 -k = 3 - -## - -[3.0, 4, 1] -comm_radius = 3.0 -seed = 1 -k = 4 - -[2.0, 4, 1] -comm_radius = 2.0 -seed = 1 -k = 4 - -[1.0, 4, 1] -comm_radius = 1.0 -seed = 1 -k = 4 - -########## - -[3.0, 1, 2] -comm_radius = 3.0 -seed = 2 -k = 1 - -[2.0, 1, 2] -comm_radius = 2.0 -seed = 2 -k = 1 - -[1.0, 1, 2] -comm_radius = 1.0 -seed = 2 -k = 1 - -## - -[3.0, 2, 2] -comm_radius = 3.0 -seed = 2 -k = 2 - -[2.0, 2, 2] -comm_radius = 2.0 -seed = 2 -k = 2 - -[1.0, 2, 2] -comm_radius = 1.0 -seed = 2 -k = 2 - -## - -[3.0, 3, 2] -comm_radius = 3.0 -seed = 2 -k = 3 - -[2.0, 3, 2] -comm_radius = 2.0 -seed = 2 -k = 3 - -[1.0, 3, 2] -comm_radius = 1.0 -seed = 2 -k = 3 - -## - -[3.0, 4, 2] -comm_radius = 3.0 -seed = 2 -k = 4 - -[2.0, 4, 2] -comm_radius = 2.0 -seed = 2 -k = 4 - -[1.0, 4, 2] -comm_radius = 1.0 -seed = 2 -k = 4 - - - - - -;[3.0, 2] -;comm_radius = 3.0 -;seed = 2 -; -;[2.0, 2] -;comm_radius = 2.0 -;seed = 2 -; -;[1.0, 2] -;comm_radius = 1.0 -;seed = 2 - - diff --git a/cfg/old/rad2.cfg b/cfg/old/rad2.cfg deleted file mode 100644 index 4ea67e7..0000000 --- a/cfg/old/rad2.cfg +++ /dev/null @@ -1,832 +0,0 @@ -[DEFAULT] - -alg = dagger - -# learning parameters -batch_size = 20 -buffer_size = 10000 -updates_per_step = 200 -seed = 11 -actor_lr = 5e-5 - -n_train_episodes = 800 -beta_coeff = 0.993 -test_interval = 40 -n_test_episodes = 20 - -# architecture parameters -k = 2 -hidden_size = 32 -gamma = 0.99 -tau = 0.5 - -# env parameters -env = FlockingRelative-v0 -v_max = 3.0 -comm_radius = 1.0 -n_agents = 100 -n_actions = 2 -n_states = 6 -debug = False -header = k, seed, comm_radius, reward - -[1, 0, 3.0] -k = 1 -seed = 0 -comm_radius = 3.0 - -[1, 0, 2.0] -k = 1 -seed = 0 -comm_radius = 2.0 - -[1, 0, 1.5] -k = 1 -seed = 0 -comm_radius = 1.5 - -[1, 0, 1.0] -k = 1 -seed = 0 -comm_radius = 1.0 - -[1, 1, 3.0] -k = 1 -seed = 1 -comm_radius = 3.0 - -[1, 1, 2.0] -k = 1 -seed = 1 -comm_radius = 2.0 - -[1, 1, 1.5] -k = 1 -seed = 1 -comm_radius = 1.5 - -[1, 1, 1.0] -k = 1 -seed = 1 -comm_radius = 1.0 - -[1, 2, 3.0] -k = 1 -seed = 2 -comm_radius = 3.0 - -[1, 2, 2.0] -k = 1 -seed = 2 -comm_radius = 2.0 - -[1, 2, 1.5] -k = 1 -seed = 2 -comm_radius = 1.5 - -[1, 2, 1.0] -k = 1 -seed = 2 -comm_radius = 1.0 - -[1, 3, 3.0] -k = 1 -seed = 3 -comm_radius = 3.0 - -[1, 3, 2.0] -k = 1 -seed = 3 -comm_radius = 2.0 - -[1, 3, 1.5] -k = 1 -seed = 3 -comm_radius = 1.5 - -[1, 3, 1.0] -k = 1 -seed = 3 -comm_radius = 1.0 - -[1, 4, 3.0] -k = 1 -seed = 4 -comm_radius = 3.0 - -[1, 4, 2.0] -k = 1 -seed = 4 -comm_radius = 2.0 - -[1, 4, 1.5] -k = 1 -seed = 4 -comm_radius = 1.5 - -[1, 4, 1.0] -k = 1 -seed = 4 -comm_radius = 1.0 - -[1, 5, 3.0] -k = 1 -seed = 5 -comm_radius = 3.0 - -[1, 5, 2.0] -k = 1 -seed = 5 -comm_radius = 2.0 - -[1, 5, 1.5] -k = 1 -seed = 5 -comm_radius = 1.5 - -[1, 5, 1.0] -k = 1 -seed = 5 -comm_radius = 1.0 - -[1, 6, 3.0] -k = 1 -seed = 6 -comm_radius = 3.0 - -[1, 6, 2.0] -k = 1 -seed = 6 -comm_radius = 2.0 - -[1, 6, 1.5] -k = 1 -seed = 6 -comm_radius = 1.5 - -[1, 6, 1.0] -k = 1 -seed = 6 -comm_radius = 1.0 - -[1, 7, 3.0] -k = 1 -seed = 7 -comm_radius = 3.0 - -[1, 7, 2.0] -k = 1 -seed = 7 -comm_radius = 2.0 - -[1, 7, 1.5] -k = 1 -seed = 7 -comm_radius = 1.5 - -[1, 7, 1.0] -k = 1 -seed = 7 -comm_radius = 1.0 - -[1, 8, 3.0] -k = 1 -seed = 8 -comm_radius = 3.0 - -[1, 8, 2.0] -k = 1 -seed = 8 -comm_radius = 2.0 - -[1, 8, 1.5] -k = 1 -seed = 8 -comm_radius = 1.5 - -[1, 8, 1.0] -k = 1 -seed = 8 -comm_radius = 1.0 - -[1, 9, 3.0] -k = 1 -seed = 9 -comm_radius = 3.0 - -[1, 9, 2.0] -k = 1 -seed = 9 -comm_radius = 2.0 - -[1, 9, 1.5] -k = 1 -seed = 9 -comm_radius = 1.5 - -[1, 9, 1.0] -k = 1 -seed = 9 -comm_radius = 1.0 - -[2, 0, 3.0] -k = 2 -seed = 0 -comm_radius = 3.0 - -[2, 0, 2.0] -k = 2 -seed = 0 -comm_radius = 2.0 - -[2, 0, 1.5] -k = 2 -seed = 0 -comm_radius = 1.5 - -[2, 0, 1.0] -k = 2 -seed = 0 -comm_radius = 1.0 - -[2, 1, 3.0] -k = 2 -seed = 1 -comm_radius = 3.0 - -[2, 1, 2.0] -k = 2 -seed = 1 -comm_radius = 2.0 - -[2, 1, 1.5] -k = 2 -seed = 1 -comm_radius = 1.5 - -[2, 1, 1.0] -k = 2 -seed = 1 -comm_radius = 1.0 - -[2, 2, 3.0] -k = 2 -seed = 2 -comm_radius = 3.0 - -[2, 2, 2.0] -k = 2 -seed = 2 -comm_radius = 2.0 - -[2, 2, 1.5] -k = 2 -seed = 2 -comm_radius = 1.5 - -[2, 2, 1.0] -k = 2 -seed = 2 -comm_radius = 1.0 - -[2, 3, 3.0] -k = 2 -seed = 3 -comm_radius = 3.0 - -[2, 3, 2.0] -k = 2 -seed = 3 -comm_radius = 2.0 - -[2, 3, 1.5] -k = 2 -seed = 3 -comm_radius = 1.5 - -[2, 3, 1.0] -k = 2 -seed = 3 -comm_radius = 1.0 - -[2, 4, 3.0] -k = 2 -seed = 4 -comm_radius = 3.0 - -[2, 4, 2.0] -k = 2 -seed = 4 -comm_radius = 2.0 - -[2, 4, 1.5] -k = 2 -seed = 4 -comm_radius = 1.5 - -[2, 4, 1.0] -k = 2 -seed = 4 -comm_radius = 1.0 - -[2, 5, 3.0] -k = 2 -seed = 5 -comm_radius = 3.0 - -[2, 5, 2.0] -k = 2 -seed = 5 -comm_radius = 2.0 - -[2, 5, 1.5] -k = 2 -seed = 5 -comm_radius = 1.5 - -[2, 5, 1.0] -k = 2 -seed = 5 -comm_radius = 1.0 - -[2, 6, 3.0] -k = 2 -seed = 6 -comm_radius = 3.0 - -[2, 6, 2.0] -k = 2 -seed = 6 -comm_radius = 2.0 - -[2, 6, 1.5] -k = 2 -seed = 6 -comm_radius = 1.5 - -[2, 6, 1.0] -k = 2 -seed = 6 -comm_radius = 1.0 - -[2, 7, 3.0] -k = 2 -seed = 7 -comm_radius = 3.0 - -[2, 7, 2.0] -k = 2 -seed = 7 -comm_radius = 2.0 - -[2, 7, 1.5] -k = 2 -seed = 7 -comm_radius = 1.5 - -[2, 7, 1.0] -k = 2 -seed = 7 -comm_radius = 1.0 - -[2, 8, 3.0] -k = 2 -seed = 8 -comm_radius = 3.0 - -[2, 8, 2.0] -k = 2 -seed = 8 -comm_radius = 2.0 - -[2, 8, 1.5] -k = 2 -seed = 8 -comm_radius = 1.5 - -[2, 8, 1.0] -k = 2 -seed = 8 -comm_radius = 1.0 - -[2, 9, 3.0] -k = 2 -seed = 9 -comm_radius = 3.0 - -[2, 9, 2.0] -k = 2 -seed = 9 -comm_radius = 2.0 - -[2, 9, 1.5] -k = 2 -seed = 9 -comm_radius = 1.5 - -[2, 9, 1.0] -k = 2 -seed = 9 -comm_radius = 1.0 - -[3, 0, 3.0] -k = 3 -seed = 0 -comm_radius = 3.0 - -[3, 0, 2.0] -k = 3 -seed = 0 -comm_radius = 2.0 - -[3, 0, 1.5] -k = 3 -seed = 0 -comm_radius = 1.5 - -[3, 0, 1.0] -k = 3 -seed = 0 -comm_radius = 1.0 - -[3, 1, 3.0] -k = 3 -seed = 1 -comm_radius = 3.0 - -[3, 1, 2.0] -k = 3 -seed = 1 -comm_radius = 2.0 - -[3, 1, 1.5] -k = 3 -seed = 1 -comm_radius = 1.5 - -[3, 1, 1.0] -k = 3 -seed = 1 -comm_radius = 1.0 - -[3, 2, 3.0] -k = 3 -seed = 2 -comm_radius = 3.0 - -[3, 2, 2.0] -k = 3 -seed = 2 -comm_radius = 2.0 - -[3, 2, 1.5] -k = 3 -seed = 2 -comm_radius = 1.5 - -[3, 2, 1.0] -k = 3 -seed = 2 -comm_radius = 1.0 - -[3, 3, 3.0] -k = 3 -seed = 3 -comm_radius = 3.0 - -[3, 3, 2.0] -k = 3 -seed = 3 -comm_radius = 2.0 - -[3, 3, 1.5] -k = 3 -seed = 3 -comm_radius = 1.5 - -[3, 3, 1.0] -k = 3 -seed = 3 -comm_radius = 1.0 - -[3, 4, 3.0] -k = 3 -seed = 4 -comm_radius = 3.0 - -[3, 4, 2.0] -k = 3 -seed = 4 -comm_radius = 2.0 - -[3, 4, 1.5] -k = 3 -seed = 4 -comm_radius = 1.5 - -[3, 4, 1.0] -k = 3 -seed = 4 -comm_radius = 1.0 - -[3, 5, 3.0] -k = 3 -seed = 5 -comm_radius = 3.0 - -[3, 5, 2.0] -k = 3 -seed = 5 -comm_radius = 2.0 - -[3, 5, 1.5] -k = 3 -seed = 5 -comm_radius = 1.5 - -[3, 5, 1.0] -k = 3 -seed = 5 -comm_radius = 1.0 - -[3, 6, 3.0] -k = 3 -seed = 6 -comm_radius = 3.0 - -[3, 6, 2.0] -k = 3 -seed = 6 -comm_radius = 2.0 - -[3, 6, 1.5] -k = 3 -seed = 6 -comm_radius = 1.5 - -[3, 6, 1.0] -k = 3 -seed = 6 -comm_radius = 1.0 - -[3, 7, 3.0] -k = 3 -seed = 7 -comm_radius = 3.0 - -[3, 7, 2.0] -k = 3 -seed = 7 -comm_radius = 2.0 - -[3, 7, 1.5] -k = 3 -seed = 7 -comm_radius = 1.5 - -[3, 7, 1.0] -k = 3 -seed = 7 -comm_radius = 1.0 - -[3, 8, 3.0] -k = 3 -seed = 8 -comm_radius = 3.0 - -[3, 8, 2.0] -k = 3 -seed = 8 -comm_radius = 2.0 - -[3, 8, 1.5] -k = 3 -seed = 8 -comm_radius = 1.5 - -[3, 8, 1.0] -k = 3 -seed = 8 -comm_radius = 1.0 - -[3, 9, 3.0] -k = 3 -seed = 9 -comm_radius = 3.0 - -[3, 9, 2.0] -k = 3 -seed = 9 -comm_radius = 2.0 - -[3, 9, 1.5] -k = 3 -seed = 9 -comm_radius = 1.5 - -[3, 9, 1.0] -k = 3 -seed = 9 -comm_radius = 1.0 - -[4, 0, 3.0] -k = 4 -seed = 0 -comm_radius = 3.0 - -[4, 0, 2.0] -k = 4 -seed = 0 -comm_radius = 2.0 - -[4, 0, 1.5] -k = 4 -seed = 0 -comm_radius = 1.5 - -[4, 0, 1.0] -k = 4 -seed = 0 -comm_radius = 1.0 - -[4, 1, 3.0] -k = 4 -seed = 1 -comm_radius = 3.0 - -[4, 1, 2.0] -k = 4 -seed = 1 -comm_radius = 2.0 - -[4, 1, 1.5] -k = 4 -seed = 1 -comm_radius = 1.5 - -[4, 1, 1.0] -k = 4 -seed = 1 -comm_radius = 1.0 - -[4, 2, 3.0] -k = 4 -seed = 2 -comm_radius = 3.0 - -[4, 2, 2.0] -k = 4 -seed = 2 -comm_radius = 2.0 - -[4, 2, 1.5] -k = 4 -seed = 2 -comm_radius = 1.5 - -[4, 2, 1.0] -k = 4 -seed = 2 -comm_radius = 1.0 - -[4, 3, 3.0] -k = 4 -seed = 3 -comm_radius = 3.0 - -[4, 3, 2.0] -k = 4 -seed = 3 -comm_radius = 2.0 - -[4, 3, 1.5] -k = 4 -seed = 3 -comm_radius = 1.5 - -[4, 3, 1.0] -k = 4 -seed = 3 -comm_radius = 1.0 - -[4, 4, 3.0] -k = 4 -seed = 4 -comm_radius = 3.0 - -[4, 4, 2.0] -k = 4 -seed = 4 -comm_radius = 2.0 - -[4, 4, 1.5] -k = 4 -seed = 4 -comm_radius = 1.5 - -[4, 4, 1.0] -k = 4 -seed = 4 -comm_radius = 1.0 - -[4, 5, 3.0] -k = 4 -seed = 5 -comm_radius = 3.0 - -[4, 5, 2.0] -k = 4 -seed = 5 -comm_radius = 2.0 - -[4, 5, 1.5] -k = 4 -seed = 5 -comm_radius = 1.5 - -[4, 5, 1.0] -k = 4 -seed = 5 -comm_radius = 1.0 - -[4, 6, 3.0] -k = 4 -seed = 6 -comm_radius = 3.0 - -[4, 6, 2.0] -k = 4 -seed = 6 -comm_radius = 2.0 - -[4, 6, 1.5] -k = 4 -seed = 6 -comm_radius = 1.5 - -[4, 6, 1.0] -k = 4 -seed = 6 -comm_radius = 1.0 - -[4, 7, 3.0] -k = 4 -seed = 7 -comm_radius = 3.0 - -[4, 7, 2.0] -k = 4 -seed = 7 -comm_radius = 2.0 - -[4, 7, 1.5] -k = 4 -seed = 7 -comm_radius = 1.5 - -[4, 7, 1.0] -k = 4 -seed = 7 -comm_radius = 1.0 - -[4, 8, 3.0] -k = 4 -seed = 8 -comm_radius = 3.0 - -[4, 8, 2.0] -k = 4 -seed = 8 -comm_radius = 2.0 - -[4, 8, 1.5] -k = 4 -seed = 8 -comm_radius = 1.5 - -[4, 8, 1.0] -k = 4 -seed = 8 -comm_radius = 1.0 - -[4, 9, 3.0] -k = 4 -seed = 9 -comm_radius = 3.0 - -[4, 9, 2.0] -k = 4 -seed = 9 -comm_radius = 2.0 - -[4, 9, 1.5] -k = 4 -seed = 9 -comm_radius = 1.5 - -[4, 9, 1.0] -k = 4 -seed = 9 -comm_radius = 1.0 - diff --git a/cfg/old/rad_baseline.cfg b/cfg/old/rad_baseline.cfg deleted file mode 100644 index 446c00d..0000000 --- a/cfg/old/rad_baseline.cfg +++ /dev/null @@ -1,434 +0,0 @@ -[DEFAULT] - -alg = baseline - -# learning parameters -batch_size = 20 -buffer_size = 10000 -updates_per_step = 200 -seed = 11 -actor_lr = 5e-5 - -n_train_episodes = 800 -beta_coeff = 0.993 -test_interval = 40 -n_test_episodes = 20 - -# architecture parameters -k = 2 -hidden_size = 32 -gamma = 0.99 -tau = 0.5 - -# env parameters -env = FlockingRelative-v0 -v_max = 3.0 -comm_radius = 1.0 -n_agents = 100 -n_actions = 2 -n_states = 6 -debug = False - - -header = comm_radius, seed, centralized, reward - -[3.0, 0, True] -comm_radius = 3.0 -seed = 0 -centralized = True - -[3.0, 0, False] -comm_radius = 3.0 -seed = 0 -centralized = False - -[3.0, 1, True] -comm_radius = 3.0 -seed = 1 -centralized = True - -[3.0, 1, False] -comm_radius = 3.0 -seed = 1 -centralized = False - -[3.0, 2, True] -comm_radius = 3.0 -seed = 2 -centralized = True - -[3.0, 2, False] -comm_radius = 3.0 -seed = 2 -centralized = False - -[3.0, 3, True] -comm_radius = 3.0 -seed = 3 -centralized = True - -[3.0, 3, False] -comm_radius = 3.0 -seed = 3 -centralized = False - -[3.0, 4, True] -comm_radius = 3.0 -seed = 4 -centralized = True - -[3.0, 4, False] -comm_radius = 3.0 -seed = 4 -centralized = False - -[3.0, 5, True] -comm_radius = 3.0 -seed = 5 -centralized = True - -[3.0, 5, False] -comm_radius = 3.0 -seed = 5 -centralized = False - -[3.0, 6, True] -comm_radius = 3.0 -seed = 6 -centralized = True - -[3.0, 6, False] -comm_radius = 3.0 -seed = 6 -centralized = False - -[3.0, 7, True] -comm_radius = 3.0 -seed = 7 -centralized = True - -[3.0, 7, False] -comm_radius = 3.0 -seed = 7 -centralized = False - -[3.0, 8, True] -comm_radius = 3.0 -seed = 8 -centralized = True - -[3.0, 8, False] -comm_radius = 3.0 -seed = 8 -centralized = False - -[3.0, 9, True] -comm_radius = 3.0 -seed = 9 -centralized = True - -[3.0, 9, False] -comm_radius = 3.0 -seed = 9 -centralized = False - -[2.0, 0, True] -comm_radius = 2.0 -seed = 0 -centralized = True - -[2.0, 0, False] -comm_radius = 2.0 -seed = 0 -centralized = False - -[2.0, 1, True] -comm_radius = 2.0 -seed = 1 -centralized = True - -[2.0, 1, False] -comm_radius = 2.0 -seed = 1 -centralized = False - -[2.0, 2, True] -comm_radius = 2.0 -seed = 2 -centralized = True - -[2.0, 2, False] -comm_radius = 2.0 -seed = 2 -centralized = False - -[2.0, 3, True] -comm_radius = 2.0 -seed = 3 -centralized = True - -[2.0, 3, False] -comm_radius = 2.0 -seed = 3 -centralized = False - -[2.0, 4, True] -comm_radius = 2.0 -seed = 4 -centralized = True - -[2.0, 4, False] -comm_radius = 2.0 -seed = 4 -centralized = False - -[2.0, 5, True] -comm_radius = 2.0 -seed = 5 -centralized = True - -[2.0, 5, False] -comm_radius = 2.0 -seed = 5 -centralized = False - -[2.0, 6, True] -comm_radius = 2.0 -seed = 6 -centralized = True - -[2.0, 6, False] -comm_radius = 2.0 -seed = 6 -centralized = False - -[2.0, 7, True] -comm_radius = 2.0 -seed = 7 -centralized = True - -[2.0, 7, False] -comm_radius = 2.0 -seed = 7 -centralized = False - -[2.0, 8, True] -comm_radius = 2.0 -seed = 8 -centralized = True - -[2.0, 8, False] -comm_radius = 2.0 -seed = 8 -centralized = False - -[2.0, 9, True] -comm_radius = 2.0 -seed = 9 -centralized = True - -[2.0, 9, False] -comm_radius = 2.0 -seed = 9 -centralized = False - -[1.5, 0, True] -comm_radius = 1.5 -seed = 0 -centralized = True - -[1.5, 0, False] -comm_radius = 1.5 -seed = 0 -centralized = False - -[1.5, 1, True] -comm_radius = 1.5 -seed = 1 -centralized = True - -[1.5, 1, False] -comm_radius = 1.5 -seed = 1 -centralized = False - -[1.5, 2, True] -comm_radius = 1.5 -seed = 2 -centralized = True - -[1.5, 2, False] -comm_radius = 1.5 -seed = 2 -centralized = False - -[1.5, 3, True] -comm_radius = 1.5 -seed = 3 -centralized = True - -[1.5, 3, False] -comm_radius = 1.5 -seed = 3 -centralized = False - -[1.5, 4, True] -comm_radius = 1.5 -seed = 4 -centralized = True - -[1.5, 4, False] -comm_radius = 1.5 -seed = 4 -centralized = False - -[1.5, 5, True] -comm_radius = 1.5 -seed = 5 -centralized = True - -[1.5, 5, False] -comm_radius = 1.5 -seed = 5 -centralized = False - -[1.5, 6, True] -comm_radius = 1.5 -seed = 6 -centralized = True - -[1.5, 6, False] -comm_radius = 1.5 -seed = 6 -centralized = False - -[1.5, 7, True] -comm_radius = 1.5 -seed = 7 -centralized = True - -[1.5, 7, False] -comm_radius = 1.5 -seed = 7 -centralized = False - -[1.5, 8, True] -comm_radius = 1.5 -seed = 8 -centralized = True - -[1.5, 8, False] -comm_radius = 1.5 -seed = 8 -centralized = False - -[1.5, 9, True] -comm_radius = 1.5 -seed = 9 -centralized = True - -[1.5, 9, False] -comm_radius = 1.5 -seed = 9 -centralized = False - -[1.0, 0, True] -comm_radius = 1.0 -seed = 0 -centralized = True - -[1.0, 0, False] -comm_radius = 1.0 -seed = 0 -centralized = False - -[1.0, 1, True] -comm_radius = 1.0 -seed = 1 -centralized = True - -[1.0, 1, False] -comm_radius = 1.0 -seed = 1 -centralized = False - -[1.0, 2, True] -comm_radius = 1.0 -seed = 2 -centralized = True - -[1.0, 2, False] -comm_radius = 1.0 -seed = 2 -centralized = False - -[1.0, 3, True] -comm_radius = 1.0 -seed = 3 -centralized = True - -[1.0, 3, False] -comm_radius = 1.0 -seed = 3 -centralized = False - -[1.0, 4, True] -comm_radius = 1.0 -seed = 4 -centralized = True - -[1.0, 4, False] -comm_radius = 1.0 -seed = 4 -centralized = False - -[1.0, 5, True] -comm_radius = 1.0 -seed = 5 -centralized = True - -[1.0, 5, False] -comm_radius = 1.0 -seed = 5 -centralized = False - -[1.0, 6, True] -comm_radius = 1.0 -seed = 6 -centralized = True - -[1.0, 6, False] -comm_radius = 1.0 -seed = 6 -centralized = False - -[1.0, 7, True] -comm_radius = 1.0 -seed = 7 -centralized = True - -[1.0, 7, False] -comm_radius = 1.0 -seed = 7 -centralized = False - -[1.0, 8, True] -comm_radius = 1.0 -seed = 8 -centralized = True - -[1.0, 8, False] -comm_radius = 1.0 -seed = 8 -centralized = False - -[1.0, 9, True] -comm_radius = 1.0 -seed = 9 -centralized = True - -[1.0, 9, False] -comm_radius = 1.0 -seed = 9 -centralized = False - diff --git a/cfg/old/vel.cfg b/cfg/old/vel.cfg deleted file mode 100644 index a64684e..0000000 --- a/cfg/old/vel.cfg +++ /dev/null @@ -1,201 +0,0 @@ -[DEFAULT] - -alg = dagger - -# learning parameters -batch_size = 20 -buffer_size = 10000 -updates_per_step = 200 -seed = 11 -actor_lr = 5e-5 - -# architecture parameters -k = 2 -hidden_size = 32 -gamma = 0.99 -tau = 0.5 - -# env parameters -env = FlockingRelative-v0 -v_max = 3.0 -comm_radius = 1.0 -n_agents = 100 -n_actions = 2 -n_states = 6 -debug = False -header = v_max, seed, k, reward - - -[3.0, 1, 1] -v_max = 3.0 -seed = 1 -k = 1 - -[2.0, 1, 1] -v_max = 2.0 -seed = 1 -k = 1 - -[1.0, 1, 1] -v_max = 1.0 -seed = 1 -k = 1 - -[0.5, 1, 1] -v_max = 0.5 -seed = 1 -k = 1 - -## - -[3.0, 2, 1] -v_max = 3.0 -seed = 1 -k = 2 - -[2.0, 2, 1] -v_max = 2.0 -seed = 1 -k = 2 - -[1.0, 2, 1] -v_max = 1.0 -seed = 1 -k = 2 - -[0.5, 2, 1] -v_max = 0.5 -seed = 1 -k = 2 - -## - -[3.0, 3, 1] -v_max = 3.0 -seed = 1 -k = 3 - -[2.0, 3, 1] -v_max = 2.0 -seed = 1 -k = 3 - -[1.0, 3, 1] -v_max = 1.0 -seed = 1 -k = 3 - -[0.5, 3, 1] -v_max = 0.5 -seed = 1 -k = 3 - -## - -[3.0, 4, 1] -v_max = 3.0 -seed = 1 -k = 4 - -[2.0, 4, 1] -v_max = 2.0 -seed = 1 -k = 4 - -[1.0, 4, 1] -v_max = 1.0 -seed = 1 -k = 4 - -[0.5, 4, 1] -v_max = 0.5 -seed = 1 -k = 4 - -########## - -[3.0, 1, 2] -v_max = 3.0 -seed = 2 -k = 1 - -[2.0, 1, 2] -v_max = 2.0 -seed = 2 -k = 1 - -[1.0, 1, 2] -v_max = 1.0 -seed = 2 -k = 1 - -[0.5, 1, 2] -v_max = 0.5 -seed = 2 -k = 1 - -## - -[3.0, 2, 2] -v_max = 3.0 -seed = 2 -k = 2 - -[2.0, 2, 2] -v_max = 2.0 -seed = 2 -k = 2 - -[1.0, 2, 2] -v_max = 1.0 -seed = 2 -k = 2 - -[0.5, 2, 2] -v_max = 0.5 -seed = 2 -k = 2 - -## - -[3.0, 3, 2] -v_max = 3.0 -seed = 2 -k = 3 - -[2.0, 3, 2] -v_max = 2.0 -seed = 2 -k = 3 - -[1.0, 3, 2] -v_max = 1.0 -seed = 2 -k = 3 - -[0.5, 3, 2] -v_max = 0.5 -seed = 2 -k = 3 - -## - -[3.0, 4, 2] -v_max = 3.0 -seed = 2 -k = 4 - -[2.0, 4, 2] -v_max = 2.0 -seed = 2 -k = 4 - -[1.0, 4, 2] -v_max = 1.0 -seed = 2 -k = 4 - -[0.5, 4, 2] -v_max = 0.5 -seed = 2 -k = 4 diff --git a/cfg/old/vel2.cfg b/cfg/old/vel2.cfg deleted file mode 100644 index 0105cf7..0000000 --- a/cfg/old/vel2.cfg +++ /dev/null @@ -1,832 +0,0 @@ -[DEFAULT] - -alg = dagger - -# learning parameters -batch_size = 20 -buffer_size = 10000 -updates_per_step = 200 -seed = 11 -actor_lr = 5e-5 - -n_train_episodes = 800 -beta_coeff = 0.993 -test_interval = 40 -n_test_episodes = 20 - -# architecture parameters -k = 2 -hidden_size = 32 -gamma = 0.99 -tau = 0.5 - -# env parameters -env = FlockingRelative-v0 -v_max = 3.0 -comm_radius = 1.0 -n_agents = 100 -n_actions = 2 -n_states = 6 -debug = False -header = v_max, k, seed, reward - -[0.5, 1, 0] -v_max = 0.5 -k = 1 -seed = 0 - -[0.5, 1, 1] -v_max = 0.5 -k = 1 -seed = 1 - -[0.5, 1, 2] -v_max = 0.5 -k = 1 -seed = 2 - -[0.5, 1, 3] -v_max = 0.5 -k = 1 -seed = 3 - -[0.5, 1, 4] -v_max = 0.5 -k = 1 -seed = 4 - -[0.5, 1, 5] -v_max = 0.5 -k = 1 -seed = 5 - -[0.5, 1, 6] -v_max = 0.5 -k = 1 -seed = 6 - -[0.5, 1, 7] -v_max = 0.5 -k = 1 -seed = 7 - -[0.5, 1, 8] -v_max = 0.5 -k = 1 -seed = 8 - -[0.5, 1, 9] -v_max = 0.5 -k = 1 -seed = 9 - -[0.5, 2, 0] -v_max = 0.5 -k = 2 -seed = 0 - -[0.5, 2, 1] -v_max = 0.5 -k = 2 -seed = 1 - -[0.5, 2, 2] -v_max = 0.5 -k = 2 -seed = 2 - -[0.5, 2, 3] -v_max = 0.5 -k = 2 -seed = 3 - -[0.5, 2, 4] -v_max = 0.5 -k = 2 -seed = 4 - -[0.5, 2, 5] -v_max = 0.5 -k = 2 -seed = 5 - -[0.5, 2, 6] -v_max = 0.5 -k = 2 -seed = 6 - -[0.5, 2, 7] -v_max = 0.5 -k = 2 -seed = 7 - -[0.5, 2, 8] -v_max = 0.5 -k = 2 -seed = 8 - -[0.5, 2, 9] -v_max = 0.5 -k = 2 -seed = 9 - -[0.5, 3, 0] -v_max = 0.5 -k = 3 -seed = 0 - -[0.5, 3, 1] -v_max = 0.5 -k = 3 -seed = 1 - -[0.5, 3, 2] -v_max = 0.5 -k = 3 -seed = 2 - -[0.5, 3, 3] -v_max = 0.5 -k = 3 -seed = 3 - -[0.5, 3, 4] -v_max = 0.5 -k = 3 -seed = 4 - -[0.5, 3, 5] -v_max = 0.5 -k = 3 -seed = 5 - -[0.5, 3, 6] -v_max = 0.5 -k = 3 -seed = 6 - -[0.5, 3, 7] -v_max = 0.5 -k = 3 -seed = 7 - -[0.5, 3, 8] -v_max = 0.5 -k = 3 -seed = 8 - -[0.5, 3, 9] -v_max = 0.5 -k = 3 -seed = 9 - -[0.5, 4, 0] -v_max = 0.5 -k = 4 -seed = 0 - -[0.5, 4, 1] -v_max = 0.5 -k = 4 -seed = 1 - -[0.5, 4, 2] -v_max = 0.5 -k = 4 -seed = 2 - -[0.5, 4, 3] -v_max = 0.5 -k = 4 -seed = 3 - -[0.5, 4, 4] -v_max = 0.5 -k = 4 -seed = 4 - -[0.5, 4, 5] -v_max = 0.5 -k = 4 -seed = 5 - -[0.5, 4, 6] -v_max = 0.5 -k = 4 -seed = 6 - -[0.5, 4, 7] -v_max = 0.5 -k = 4 -seed = 7 - -[0.5, 4, 8] -v_max = 0.5 -k = 4 -seed = 8 - -[0.5, 4, 9] -v_max = 0.5 -k = 4 -seed = 9 - -[1.5, 1, 0] -v_max = 1.5 -k = 1 -seed = 0 - -[1.5, 1, 1] -v_max = 1.5 -k = 1 -seed = 1 - -[1.5, 1, 2] -v_max = 1.5 -k = 1 -seed = 2 - -[1.5, 1, 3] -v_max = 1.5 -k = 1 -seed = 3 - -[1.5, 1, 4] -v_max = 1.5 -k = 1 -seed = 4 - -[1.5, 1, 5] -v_max = 1.5 -k = 1 -seed = 5 - -[1.5, 1, 6] -v_max = 1.5 -k = 1 -seed = 6 - -[1.5, 1, 7] -v_max = 1.5 -k = 1 -seed = 7 - -[1.5, 1, 8] -v_max = 1.5 -k = 1 -seed = 8 - -[1.5, 1, 9] -v_max = 1.5 -k = 1 -seed = 9 - -[1.5, 2, 0] -v_max = 1.5 -k = 2 -seed = 0 - -[1.5, 2, 1] -v_max = 1.5 -k = 2 -seed = 1 - -[1.5, 2, 2] -v_max = 1.5 -k = 2 -seed = 2 - -[1.5, 2, 3] -v_max = 1.5 -k = 2 -seed = 3 - -[1.5, 2, 4] -v_max = 1.5 -k = 2 -seed = 4 - -[1.5, 2, 5] -v_max = 1.5 -k = 2 -seed = 5 - -[1.5, 2, 6] -v_max = 1.5 -k = 2 -seed = 6 - -[1.5, 2, 7] -v_max = 1.5 -k = 2 -seed = 7 - -[1.5, 2, 8] -v_max = 1.5 -k = 2 -seed = 8 - -[1.5, 2, 9] -v_max = 1.5 -k = 2 -seed = 9 - -[1.5, 3, 0] -v_max = 1.5 -k = 3 -seed = 0 - -[1.5, 3, 1] -v_max = 1.5 -k = 3 -seed = 1 - -[1.5, 3, 2] -v_max = 1.5 -k = 3 -seed = 2 - -[1.5, 3, 3] -v_max = 1.5 -k = 3 -seed = 3 - -[1.5, 3, 4] -v_max = 1.5 -k = 3 -seed = 4 - -[1.5, 3, 5] -v_max = 1.5 -k = 3 -seed = 5 - -[1.5, 3, 6] -v_max = 1.5 -k = 3 -seed = 6 - -[1.5, 3, 7] -v_max = 1.5 -k = 3 -seed = 7 - -[1.5, 3, 8] -v_max = 1.5 -k = 3 -seed = 8 - -[1.5, 3, 9] -v_max = 1.5 -k = 3 -seed = 9 - -[1.5, 4, 0] -v_max = 1.5 -k = 4 -seed = 0 - -[1.5, 4, 1] -v_max = 1.5 -k = 4 -seed = 1 - -[1.5, 4, 2] -v_max = 1.5 -k = 4 -seed = 2 - -[1.5, 4, 3] -v_max = 1.5 -k = 4 -seed = 3 - -[1.5, 4, 4] -v_max = 1.5 -k = 4 -seed = 4 - -[1.5, 4, 5] -v_max = 1.5 -k = 4 -seed = 5 - -[1.5, 4, 6] -v_max = 1.5 -k = 4 -seed = 6 - -[1.5, 4, 7] -v_max = 1.5 -k = 4 -seed = 7 - -[1.5, 4, 8] -v_max = 1.5 -k = 4 -seed = 8 - -[1.5, 4, 9] -v_max = 1.5 -k = 4 -seed = 9 - -[2.5, 1, 0] -v_max = 2.5 -k = 1 -seed = 0 - -[2.5, 1, 1] -v_max = 2.5 -k = 1 -seed = 1 - -[2.5, 1, 2] -v_max = 2.5 -k = 1 -seed = 2 - -[2.5, 1, 3] -v_max = 2.5 -k = 1 -seed = 3 - -[2.5, 1, 4] -v_max = 2.5 -k = 1 -seed = 4 - -[2.5, 1, 5] -v_max = 2.5 -k = 1 -seed = 5 - -[2.5, 1, 6] -v_max = 2.5 -k = 1 -seed = 6 - -[2.5, 1, 7] -v_max = 2.5 -k = 1 -seed = 7 - -[2.5, 1, 8] -v_max = 2.5 -k = 1 -seed = 8 - -[2.5, 1, 9] -v_max = 2.5 -k = 1 -seed = 9 - -[2.5, 2, 0] -v_max = 2.5 -k = 2 -seed = 0 - -[2.5, 2, 1] -v_max = 2.5 -k = 2 -seed = 1 - -[2.5, 2, 2] -v_max = 2.5 -k = 2 -seed = 2 - -[2.5, 2, 3] -v_max = 2.5 -k = 2 -seed = 3 - -[2.5, 2, 4] -v_max = 2.5 -k = 2 -seed = 4 - -[2.5, 2, 5] -v_max = 2.5 -k = 2 -seed = 5 - -[2.5, 2, 6] -v_max = 2.5 -k = 2 -seed = 6 - -[2.5, 2, 7] -v_max = 2.5 -k = 2 -seed = 7 - -[2.5, 2, 8] -v_max = 2.5 -k = 2 -seed = 8 - -[2.5, 2, 9] -v_max = 2.5 -k = 2 -seed = 9 - -[2.5, 3, 0] -v_max = 2.5 -k = 3 -seed = 0 - -[2.5, 3, 1] -v_max = 2.5 -k = 3 -seed = 1 - -[2.5, 3, 2] -v_max = 2.5 -k = 3 -seed = 2 - -[2.5, 3, 3] -v_max = 2.5 -k = 3 -seed = 3 - -[2.5, 3, 4] -v_max = 2.5 -k = 3 -seed = 4 - -[2.5, 3, 5] -v_max = 2.5 -k = 3 -seed = 5 - -[2.5, 3, 6] -v_max = 2.5 -k = 3 -seed = 6 - -[2.5, 3, 7] -v_max = 2.5 -k = 3 -seed = 7 - -[2.5, 3, 8] -v_max = 2.5 -k = 3 -seed = 8 - -[2.5, 3, 9] -v_max = 2.5 -k = 3 -seed = 9 - -[2.5, 4, 0] -v_max = 2.5 -k = 4 -seed = 0 - -[2.5, 4, 1] -v_max = 2.5 -k = 4 -seed = 1 - -[2.5, 4, 2] -v_max = 2.5 -k = 4 -seed = 2 - -[2.5, 4, 3] -v_max = 2.5 -k = 4 -seed = 3 - -[2.5, 4, 4] -v_max = 2.5 -k = 4 -seed = 4 - -[2.5, 4, 5] -v_max = 2.5 -k = 4 -seed = 5 - -[2.5, 4, 6] -v_max = 2.5 -k = 4 -seed = 6 - -[2.5, 4, 7] -v_max = 2.5 -k = 4 -seed = 7 - -[2.5, 4, 8] -v_max = 2.5 -k = 4 -seed = 8 - -[2.5, 4, 9] -v_max = 2.5 -k = 4 -seed = 9 - -[3.5, 1, 0] -v_max = 3.5 -k = 1 -seed = 0 - -[3.5, 1, 1] -v_max = 3.5 -k = 1 -seed = 1 - -[3.5, 1, 2] -v_max = 3.5 -k = 1 -seed = 2 - -[3.5, 1, 3] -v_max = 3.5 -k = 1 -seed = 3 - -[3.5, 1, 4] -v_max = 3.5 -k = 1 -seed = 4 - -[3.5, 1, 5] -v_max = 3.5 -k = 1 -seed = 5 - -[3.5, 1, 6] -v_max = 3.5 -k = 1 -seed = 6 - -[3.5, 1, 7] -v_max = 3.5 -k = 1 -seed = 7 - -[3.5, 1, 8] -v_max = 3.5 -k = 1 -seed = 8 - -[3.5, 1, 9] -v_max = 3.5 -k = 1 -seed = 9 - -[3.5, 2, 0] -v_max = 3.5 -k = 2 -seed = 0 - -[3.5, 2, 1] -v_max = 3.5 -k = 2 -seed = 1 - -[3.5, 2, 2] -v_max = 3.5 -k = 2 -seed = 2 - -[3.5, 2, 3] -v_max = 3.5 -k = 2 -seed = 3 - -[3.5, 2, 4] -v_max = 3.5 -k = 2 -seed = 4 - -[3.5, 2, 5] -v_max = 3.5 -k = 2 -seed = 5 - -[3.5, 2, 6] -v_max = 3.5 -k = 2 -seed = 6 - -[3.5, 2, 7] -v_max = 3.5 -k = 2 -seed = 7 - -[3.5, 2, 8] -v_max = 3.5 -k = 2 -seed = 8 - -[3.5, 2, 9] -v_max = 3.5 -k = 2 -seed = 9 - -[3.5, 3, 0] -v_max = 3.5 -k = 3 -seed = 0 - -[3.5, 3, 1] -v_max = 3.5 -k = 3 -seed = 1 - -[3.5, 3, 2] -v_max = 3.5 -k = 3 -seed = 2 - -[3.5, 3, 3] -v_max = 3.5 -k = 3 -seed = 3 - -[3.5, 3, 4] -v_max = 3.5 -k = 3 -seed = 4 - -[3.5, 3, 5] -v_max = 3.5 -k = 3 -seed = 5 - -[3.5, 3, 6] -v_max = 3.5 -k = 3 -seed = 6 - -[3.5, 3, 7] -v_max = 3.5 -k = 3 -seed = 7 - -[3.5, 3, 8] -v_max = 3.5 -k = 3 -seed = 8 - -[3.5, 3, 9] -v_max = 3.5 -k = 3 -seed = 9 - -[3.5, 4, 0] -v_max = 3.5 -k = 4 -seed = 0 - -[3.5, 4, 1] -v_max = 3.5 -k = 4 -seed = 1 - -[3.5, 4, 2] -v_max = 3.5 -k = 4 -seed = 2 - -[3.5, 4, 3] -v_max = 3.5 -k = 4 -seed = 3 - -[3.5, 4, 4] -v_max = 3.5 -k = 4 -seed = 4 - -[3.5, 4, 5] -v_max = 3.5 -k = 4 -seed = 5 - -[3.5, 4, 6] -v_max = 3.5 -k = 4 -seed = 6 - -[3.5, 4, 7] -v_max = 3.5 -k = 4 -seed = 7 - -[3.5, 4, 8] -v_max = 3.5 -k = 4 -seed = 8 - -[3.5, 4, 9] -v_max = 3.5 -k = 4 -seed = 9 - diff --git a/cfg/old/vel_baseline.cfg b/cfg/old/vel_baseline.cfg deleted file mode 100644 index 48a16d6..0000000 --- a/cfg/old/vel_baseline.cfg +++ /dev/null @@ -1,434 +0,0 @@ -[DEFAULT] - -alg = baseline - -# learning parameters -batch_size = 20 -buffer_size = 10000 -updates_per_step = 200 -seed = 11 -actor_lr = 5e-5 - -n_train_episodes = 800 -beta_coeff = 0.993 -test_interval = 40 -n_test_episodes = 20 - -# architecture parameters -k = 2 -hidden_size = 32 -gamma = 0.99 -tau = 0.5 - -# env parameters -env = FlockingRelative-v0 -v_max = 3.0 -comm_radius = 1.0 -n_agents = 100 -n_actions = 2 -n_states = 6 -debug = False - - -header = v_max, seed, centralized, reward - -[0.5, 0, True] -v_max = 0.5 -seed = 0 -centralized = True - -[0.5, 0, False] -v_max = 0.5 -seed = 0 -centralized = False - -[0.5, 1, True] -v_max = 0.5 -seed = 1 -centralized = True - -[0.5, 1, False] -v_max = 0.5 -seed = 1 -centralized = False - -[0.5, 2, True] -v_max = 0.5 -seed = 2 -centralized = True - -[0.5, 2, False] -v_max = 0.5 -seed = 2 -centralized = False - -[0.5, 3, True] -v_max = 0.5 -seed = 3 -centralized = True - -[0.5, 3, False] -v_max = 0.5 -seed = 3 -centralized = False - -[0.5, 4, True] -v_max = 0.5 -seed = 4 -centralized = True - -[0.5, 4, False] -v_max = 0.5 -seed = 4 -centralized = False - -[0.5, 5, True] -v_max = 0.5 -seed = 5 -centralized = True - -[0.5, 5, False] -v_max = 0.5 -seed = 5 -centralized = False - -[0.5, 6, True] -v_max = 0.5 -seed = 6 -centralized = True - -[0.5, 6, False] -v_max = 0.5 -seed = 6 -centralized = False - -[0.5, 7, True] -v_max = 0.5 -seed = 7 -centralized = True - -[0.5, 7, False] -v_max = 0.5 -seed = 7 -centralized = False - -[0.5, 8, True] -v_max = 0.5 -seed = 8 -centralized = True - -[0.5, 8, False] -v_max = 0.5 -seed = 8 -centralized = False - -[0.5, 9, True] -v_max = 0.5 -seed = 9 -centralized = True - -[0.5, 9, False] -v_max = 0.5 -seed = 9 -centralized = False - -[1.5, 0, True] -v_max = 1.5 -seed = 0 -centralized = True - -[1.5, 0, False] -v_max = 1.5 -seed = 0 -centralized = False - -[1.5, 1, True] -v_max = 1.5 -seed = 1 -centralized = True - -[1.5, 1, False] -v_max = 1.5 -seed = 1 -centralized = False - -[1.5, 2, True] -v_max = 1.5 -seed = 2 -centralized = True - -[1.5, 2, False] -v_max = 1.5 -seed = 2 -centralized = False - -[1.5, 3, True] -v_max = 1.5 -seed = 3 -centralized = True - -[1.5, 3, False] -v_max = 1.5 -seed = 3 -centralized = False - -[1.5, 4, True] -v_max = 1.5 -seed = 4 -centralized = True - -[1.5, 4, False] -v_max = 1.5 -seed = 4 -centralized = False - -[1.5, 5, True] -v_max = 1.5 -seed = 5 -centralized = True - -[1.5, 5, False] -v_max = 1.5 -seed = 5 -centralized = False - -[1.5, 6, True] -v_max = 1.5 -seed = 6 -centralized = True - -[1.5, 6, False] -v_max = 1.5 -seed = 6 -centralized = False - -[1.5, 7, True] -v_max = 1.5 -seed = 7 -centralized = True - -[1.5, 7, False] -v_max = 1.5 -seed = 7 -centralized = False - -[1.5, 8, True] -v_max = 1.5 -seed = 8 -centralized = True - -[1.5, 8, False] -v_max = 1.5 -seed = 8 -centralized = False - -[1.5, 9, True] -v_max = 1.5 -seed = 9 -centralized = True - -[1.5, 9, False] -v_max = 1.5 -seed = 9 -centralized = False - -[2.5, 0, True] -v_max = 2.5 -seed = 0 -centralized = True - -[2.5, 0, False] -v_max = 2.5 -seed = 0 -centralized = False - -[2.5, 1, True] -v_max = 2.5 -seed = 1 -centralized = True - -[2.5, 1, False] -v_max = 2.5 -seed = 1 -centralized = False - -[2.5, 2, True] -v_max = 2.5 -seed = 2 -centralized = True - -[2.5, 2, False] -v_max = 2.5 -seed = 2 -centralized = False - -[2.5, 3, True] -v_max = 2.5 -seed = 3 -centralized = True - -[2.5, 3, False] -v_max = 2.5 -seed = 3 -centralized = False - -[2.5, 4, True] -v_max = 2.5 -seed = 4 -centralized = True - -[2.5, 4, False] -v_max = 2.5 -seed = 4 -centralized = False - -[2.5, 5, True] -v_max = 2.5 -seed = 5 -centralized = True - -[2.5, 5, False] -v_max = 2.5 -seed = 5 -centralized = False - -[2.5, 6, True] -v_max = 2.5 -seed = 6 -centralized = True - -[2.5, 6, False] -v_max = 2.5 -seed = 6 -centralized = False - -[2.5, 7, True] -v_max = 2.5 -seed = 7 -centralized = True - -[2.5, 7, False] -v_max = 2.5 -seed = 7 -centralized = False - -[2.5, 8, True] -v_max = 2.5 -seed = 8 -centralized = True - -[2.5, 8, False] -v_max = 2.5 -seed = 8 -centralized = False - -[2.5, 9, True] -v_max = 2.5 -seed = 9 -centralized = True - -[2.5, 9, False] -v_max = 2.5 -seed = 9 -centralized = False - -[3.5, 0, True] -v_max = 3.5 -seed = 0 -centralized = True - -[3.5, 0, False] -v_max = 3.5 -seed = 0 -centralized = False - -[3.5, 1, True] -v_max = 3.5 -seed = 1 -centralized = True - -[3.5, 1, False] -v_max = 3.5 -seed = 1 -centralized = False - -[3.5, 2, True] -v_max = 3.5 -seed = 2 -centralized = True - -[3.5, 2, False] -v_max = 3.5 -seed = 2 -centralized = False - -[3.5, 3, True] -v_max = 3.5 -seed = 3 -centralized = True - -[3.5, 3, False] -v_max = 3.5 -seed = 3 -centralized = False - -[3.5, 4, True] -v_max = 3.5 -seed = 4 -centralized = True - -[3.5, 4, False] -v_max = 3.5 -seed = 4 -centralized = False - -[3.5, 5, True] -v_max = 3.5 -seed = 5 -centralized = True - -[3.5, 5, False] -v_max = 3.5 -seed = 5 -centralized = False - -[3.5, 6, True] -v_max = 3.5 -seed = 6 -centralized = True - -[3.5, 6, False] -v_max = 3.5 -seed = 6 -centralized = False - -[3.5, 7, True] -v_max = 3.5 -seed = 7 -centralized = True - -[3.5, 7, False] -v_max = 3.5 -seed = 7 -centralized = False - -[3.5, 8, True] -v_max = 3.5 -seed = 8 -centralized = True - -[3.5, 8, False] -v_max = 3.5 -seed = 8 -centralized = False - -[3.5, 9, True] -v_max = 3.5 -seed = 9 -centralized = True - -[3.5, 9, False] -v_max = 3.5 -seed = 9 -centralized = False - diff --git a/learner/gnn_cloning.py b/learner/gnn_cloning.py index 00898f9..69e1353 100644 --- a/learner/gnn_cloning.py +++ b/learner/gnn_cloning.py @@ -105,7 +105,7 @@ def save_model(self, env_name, suffix="", actor_path=None): os.makedirs('models/') if actor_path is None: - actor_path = "models/ddpg_actor_{}_{}".format(env_name, suffix) + actor_path = "models/actor_{}_{}".format(env_name, suffix) print('Saving model to {}'.format(actor_path)) torch.save(self.actor.state_dict(), actor_path) diff --git a/learner/gnn_dagger.py b/learner/gnn_dagger.py index 4a1e14a..8e310be 100644 --- a/learner/gnn_dagger.py +++ b/learner/gnn_dagger.py @@ -107,7 +107,7 @@ def save_model(self, env_name, suffix="", actor_path=None): os.makedirs('models/') if actor_path is None: - actor_path = "models/ddpg_actor_{}_{}".format(env_name, suffix) + actor_path = "models/actor_{}_{}".format(env_name, suffix) print('Saving model to {}'.format(actor_path)) torch.save(self.actor.state_dict(), actor_path) diff --git a/models/actor_FlockingRelative-v0_dagger_k3 b/models/actor_FlockingRelative-v0_dagger_k3 new file mode 100644 index 0000000000000000000000000000000000000000..7e1941881e15e884caab0c56a84699e1266bba8c GIT binary patch literal 7990 zcmZvgd0b83_y3zz8c-C`tVuMN=$^f9GDJzq6d^^aq`D2pqCsgO(WKI#P)ZV+x@WHw zB9SSQH)TlXkRe3sTkp5e@A3P-Ki~7meeAW@Ua#}qXRW>OJ|{{nNhYB8z^ook|56tf z<0vtwH{t`okpd9|HyIHTk+99d-X5WTTf9RBqLBhI12^%3PQSn~fw-fDo5VmnASiIH zKw`icGGK)I28DzPhB%6ayNwtygF}MCf>s9wcx>6i6k658QM0z{mja)nNkzLuH&p)_8|_ubJVuI!s0&GQm|KYT!0%AaHe1;0BKX z&&>mZ#-_#_z5Ukuh6%)0hXsYK_BHkh5AzEMmGKDi_6qk4SmP1q9T*xE;<3S8Mj&or z;3hRlI55(4t+znppAeZ@0YRQ&3;yYp5e#vY9vEY7YGyfgin*mg^6%NM4h|PcIoXQM z3lm7scN8&j)O3_|6n7L26AU$QR2*y#wq+6r{IaeBIXA`s!u0a<3>C=x4lwvi`2KT` z$x}_IOtCaI6%6y0a?~6M9X=rKs2L|1;V2p>P#6?bbQO&JuQ{9jXU?PkB{$V>Z8dqawYfm`531T9 zRP{krjX^O@SAo`lp_=~(Rr_CZ7UtGw=B9%219G+l!8(7Fiv7bECeZ!=Nc9HA^<4!9 z|3zy459x$|Nn4v*n+_z-a6o!6UlRvO#o`2$21$(u#f)7ACgG6+Q@5c5S@Kx#9p<^l zGt5(9<~vwp<_2y9y&?jONP*?w)$*UZwQ>{cUcuK}7;CTN#K~hdJh`qy=NS5+`Ui0+TarR}I(cl$PX&0AUe4GB z7oLxteu1riZ;80MnZDK?gZF%U$-FBP!q^x!s$$g2s@2V+IkAf9^k_YXuI?f$&q&}; z2jRJo2LeET_y%_8(g|!8z9rMEEO6}hF*xkNBx?INg@_2hI*)(yTowPl@qT^;*~9Pu z6wNJMo61R(GHzvc4!=x)9{2R<8NTh>)11!898M}Dha1gMd~09Cmk*2KbezlhN1haN zNdc8yS@ceRq2)2Y`^8Gm=~NYelw27nzo(3oT~*D;-V$#A&U~&!v78@Tk;G}OKgM+x z74n_O=Wth>O!51?LwwJ>M>$?Z3IE%w6Z|Jyo4K&J)%;+alYBqRVE%>fIw)LI#I5fs znNA=-yb~85BOco7yI)Z)>@1vm9u53<{!J*>@<5vb~QP+_Bze= z5+fpSmht3nMA2kPO`5gq7=1n`g06}gD%{Z@$@-kkCUs5z%;X%(8raPyrg38A!V*im z^!j^N&rY3|AN3+`ehus|MO7-kp`VHNbs{C3Fie1F?m%KdM78qNWo(qH_D5DR(uNjIela5ua2WubNA3Tyv)nc@J`kj}iokRl?0B2-Y z;^3@Tg(ybU0TYN2-R8=G^KHh|vpW60{>4 zHyb|&CPf+MJl4g`OO{wL%?L{MFV%FJ@YsyZZ(!M;iV+68!1i<}I28rZ{W>h}$>6as zPW6&JyGi6>mKL?QIST!a&FGwG0QOt6VgJxL?6T^s>HS&^t9AWoxMKkbwv6VC^^U`* zmwRaumkg&QN0L3);^4W4C~JObJ#*z&GORrwMvlw4z!29`YCF#WkMs=GKW_~PPI!?E z-wo+j5W_nK;`~XQb8*?UFH|ae2&$@=LGZ{16m7_bt@8`$*{z*umuUm(^G*?Khje5D zqZrv`(fGmBfIs468y@jd!k>9bXiKh8!$2S0zfOdd#XG=^V-0BM6-av-9JW6rr?LZuzf_P@^k5UszL(h}?4q(-C^ zr%P80fBscKPaL%a$&a4kQ)*SC=w40L9*0m9`!?c;LipkG8QqhYLW7GsbXpJLy?Lw8 zw_~9L=rOSK((asfM9395wLaf%z>pZtl=w;QgzKx;f_x zBQDp|*&Tc?RJxzny{s56XEng@n(6e4X#yCfX2K4?%lO7J1$Jn*;@B=-j(6V!>}S|8 zdQ}{7lJPWH$AK996^@Pj>+p@bAM}k+MIXTKK1UpYv)VJcd(VdAsx9^O>di!uE=nfS&2;XuLSZ^no=>BZ!pT$pCsP5!WCUHxZ##LDf6wT4}{p+}9#!dehv7 zc88q8xv6DjLP`kgOPcd1FR{U(TgxytumgXqDUwxVcv#$aiD}XFf)tp9Ir^UT!PRX{ z&J%6;JYg+7?G^?3yRz7~K!e0e=P{p*F2XMJ-`E>e05^Mduq)jRq@PaUhORqI6&|>5I9z z+#nhCib^riM2DYUc!y1v8HpM{-@?ha7eHXV21R^#LzdPxR^;qPHhq&dPP9|!ceb~Y zug`vxfJYLT7nzCS#}xQ%L_A%tDNl!tbbzLBcktf*hjdDxHN3rd74F?Tz!+;TgTlFT z@b2R%u2cFM6ehIbQuTIx&?^HrFS4LwvlV=eSb~k7Pf64-Z!*3y8uxH9pwc#<=J{2l z{If}vS^Jd+?Jl6XDH71c}b*8zDm6FmY@g*ci)nXX0o&21ZLuN$?$W(kK@ld#8!4%@EpboJ|i{UTfVNjhUhHm5L35TqVhuYBvz%ajY zMcD(gGPa(CJZxhUJy={7a0$4X23R+0I2r%48uC`ItvNDZA9^CCz|7SP`xN)W^5!wz z%)4g%i`|vP-%bbM#17uI-KS`fZVfr$8%W)ro+TECM`7M4Eim;f#EOSGXrEbWLtZ)1 zW;1=bA0dvq#s+ISURo1dvS*viMhRHpGT3gpMbNFg8dT#x6{w ze#duX@vIy4SjZ?YQTHM!T}h=Q{5Tpm<`cTD8-mq-wosiP0Sj-$;Ix)&C?lw$ecv}> zh|j=Y9F~h}Iy)d#(HSFDVj#M6B-&3x^uHT}1!DI}blPli4}L_86o7Yy_-M z$VB5g^|VgP4BN8ag4?JZy6@&3JUv^9uP|{8S-AQJTnV2GD$>Eg`#cI%9u2`ORmKou z^%SZDw?et?PBKzsGJit%8(>_rNcox!x+D4~IaIS3b=WVUwILQp_~*j;531ZcuO-zNy9F2Sz3DXeOpDrX-xiVIe` zM;>cvq44ur_`)lqa*I2O#IOJuFLw~dElENa~>~CVEa!w{`EZpIC9^4vgK5w@WM78yiUr*<}wAo z&zA=B$ypzEYo)RB?Vrf^S54rO6KNJkS6AYmhIZ8L+KrvF$KmI;IJ$500o<~! z7R@Kfat8T#X~^LKBr9e4X+;qj7gq){6Gn2M%uka8hI3K*QWetbJ}i#7jDcaVNv7!t z_H&FMK3uaK6r00firH1z8C8sCW*=~Aoftlss)ZkFLRfgPolM)632oXeQA#NU{jQ}T zzbTbHD53{0y=&kO-=43vK#|`%JfH3Q)k%hEH=yQ8Ka|L7BwyCNg*Qd#n4(}Kd~@*> zX%Ok7T0YIhCP;`M?HqN7_7N40Y_RFNf?ItP&{^{*eoS;Fa?@T4 zjaDjQ{J{SotQ6wp`bOLrEQ?F0Rj~#4?6@C~3eoR!6R9mV-D`e zhLS8isiDn(_v->EiM&Snj#}JXJc3B4zJ>nT2QlvKPDt>KB^u|S!>blMxKamm$P%y>o7q{hCG-31ycKRzzFT3Rips7wKvcu zR_d59=7>}8+(ym!id;&F72<0*oauB9-OioG?Jm=~>L*HgW6CtNzch@-+zACUnQb`n z{v}vZIfPU2RKmCrG2ymm#75l|+)ZO}V^j+4?+Bv@E>Fhlju<*|=2560BLg>n>v1|V z2XTtYDoV@TVC<7Ja$^l-O9jGa-GL$H`utkG$8aWM z7^K{oOv(=Xm7vK(%}m^mrW! zW~RNo!nAIX4Ec-7U7Ufw$CFUviaVh%k707)38;9`Nr z#$+nqOT58re4WhXi;d&N&YdG4;`U+r^|Q2PWy!0s=U!`gb!<%lAXFHY1>KD@dQ(JQI=i<#cXE}>0-*RBXu_!!#?*v5e=tp71 zFZv=A={mNLjyv>0$d%jSLc1NTZDQLSAz-Ui#h+sSdk$&!a$>82sk!v#2le3 z49j1R);4K$e}p65lz4?K=r+dhQsHQ#TZGdK>xo0XGIvp93cR-07+BMDp|4Ps>NHoJ zpF^{GH^lVewQ~q9R)|Lr2QlIx70*<3_mZS>MK$xi&(>@bH(}K~4CsrowStWF-7T3w7FJeLhoqNeTP%Q=<_w7VHk~XoNu-k|j)uB- zXXucpVff6doLzmr1)F*pGFx1o3A807XTvzQQ~EevJo<-?UHdY6uVp7IVzClyv%jG7 zvjkfBwG}m+UJ%V5BglUKgXOPNfo;B7?0AzN*uK&Vex=khix+2*-&SE%VfHIt?VJ&C zJlYu!oNl5|3#UMT*ka)PFOrUreUN)kpSGu)G21QUsF95`?c0mAvM&O&RrM)EE0J9< zve2$Si)Z=CfUPfSscCp^NWVmcLjJ~l=unLVpNaM$PwYYY=n}f$#EHGxZ_mr05DOFY zvzgKDN`#pCLsDfXr6Ws->ewur6&_0x7M4QV@aMuOS#?a7a|6i4NRd4+8LIH$u<*qF z9qgkwG4%b;okVMa160(l{Z^Y-%Swb_*D>sT?^r7RW*104RfC&v_A_Pqgj{z|hSC71 zbB>4GY+fr_kVTv_Ze8C>PP$CSnfcn_y+NPoYDd6Qg)i*RInSAK!bwnmC5*PJmfF1D~a_vJrN~N+y~0bju-D zYvho1olIqq_@dRBcqX_*ku4QJB0L&X&zo|^6;{N1)7)JdO!GDkc6n7f;l18I)nwit zo4XN_ym?zpSmm6Pv`|5dczq}$BbSYWsQb_1hD{B7r!d3jP4V=aPSy?re+b1(CRdE(_BCgyE{{VsR(?2Lz%bOrP;71*}hTfU*q|@Z)Ya1Y`nr)vDrMwwEK3Nf{(Ob2q^+v7qhn1F9pY z(`mDspyrt&wlwv!ms^jMdvejBe_|$i+EYZ^$NR&~%w)KAN`fpn4-y_#vjz-(@ zVq+PFwYW#W=|z>Q5syvsUe-~L5sW7aJr!w zPgbLtF0XLF;?Xr2ad#!P7JzWW>o2;PFZx5XLm_pzF z`o)?|O``eXjZCw>6Hh(I0-1{sh_c9WqU5-k91Bv!z)c}SAJ1%zD7r_}R;zIK*`}OH z)Oj=xDkjkv<`XB*kZ$?G5cTvUSUNt6T_E3w3#?zE|BNYUxb80f5u1QTEA5Et7c~-f zqKp~0(-x(cQM&AWJZ|1l%T(1MyY!AS^uCqmg|x@>YI9c!D>wJjcfMkD`lg3u`2umc zacC>*{paG*1D(bNR4$k~GG<2NBW67Wzfo6Xx#qMV@IPN^l2I@>LI}UFv7|8Hl4s z;$gDvW)4w!<9#k6Bb^-9ze=+>GgdLY2jvj4=f1@ zAdfd_k@M%qk?IR`arB%RocQAs4l}aB_p%D8WBh>5x}i%Lkr^cBj2z>oVMfm^u0o&R zdT8`w5&b%OGBfw`F02ynsWFPHWV0TPVKuIDWaC&7klTHUB(5~z7M@XJrmp*qzh;Q> zB4gGO1!-$^+N_7_f%Cbx0$omGc^OStKMgB2JyFkX9K9uBh!($Jl9!jWVf#A~{)Z*f zd{Ux=P{G4}SHICG?UJOZ{WEVtnLqUrEu@$F(&+tNGSKTT4$r+TLAmGvOmIEG&Y`oI zAAQ-RZCf%Wohz{GuN2I6Rs$>*hYFc?@_?14iQB7rXM#Dlhi#;9%N*dHjwlX!vXf0` iG4;-0k-t(^V;2F$o~KU31}(+ literal 0 HcmV?d00001 diff --git a/test_3pane.py b/results/figure_code/test_3pane.py similarity index 100% rename from test_3pane.py rename to results/figure_code/test_3pane.py diff --git a/test_one_traj.py b/results/figure_code/test_one_traj.py similarity index 100% rename from test_one_traj.py rename to results/figure_code/test_one_traj.py diff --git a/test_one_traj2.py b/results/figure_code/test_one_traj2.py similarity index 100% rename from test_one_traj2.py rename to results/figure_code/test_one_traj2.py diff --git a/test_model.py b/test_model.py index 3c3653a..eef609c 100644 --- a/test_model.py +++ b/test_model.py @@ -11,12 +11,10 @@ from learner.gnn_dagger import DAGGER -def test(args): +def test(args, actor_path, render=True): # initialize gym env env_name = args.get('env') - env_name = "FlockingAirsimAccel-v0" env = gym.make(env_name) - if isinstance(env.env, gym_flock.envs.FlockingRelativeEnv): env.env.params_from_cfg(args) @@ -31,10 +29,6 @@ def test(args): device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu") learner = DAGGER(device, args) n_test_episodes = args.getint('n_test_episodes') - # actor_path = 'models/ddpg_actor_FlockingRelative-v0_k3' - # actor_path = 'models/ddpg_actor_FlockingRelative-v0_k3' - actor_path = 'models/ddpg_actor_FlockingStochastic-v0_stoch2' - learner.load_model(actor_path, device) for _ in range(n_test_episodes): @@ -47,7 +41,8 @@ def test(args): next_state = MultiAgentStateWithDelay(device, args, next_state, prev_state=state) episode_reward += reward state = next_state - #env.render() + if render: + env.render() print(episode_reward) env.close() @@ -59,6 +54,7 @@ def main(): config.read(config_file) printed_header = False + actor_path = 'models/actor_FlockingRelative-v0_dagger_k3' if config.sections(): for section_name in config.sections(): @@ -66,9 +62,9 @@ def main(): print(config[section_name].get('header')) printed_header = True - test(config[section_name]) + test(config[section_name], actor_path) else: - test(config[config.default_section]) + test(config[config.default_section], actor_path)