config files for all experiments

rfal · rfal · commit 61cab8f1bbd9 · 2022-06-19T22:47:39.000-04:00
diff --git a/.gitignore b/.gitignore
@@ -2,3 +2,4 @@ carla_data
 Baselines
 carle_gym/carle.egg-info
 __pycache__
+scripts
diff --git a/config/config_A2C_Town01_mlp.json b/config/config_A2C_Town01_mlp.json
@@ -1,6 +1,6 @@
 {
   "base":{
-    "seed": [0,1,2,3,4],
+    "seed": [0],
     "eval_freq": 10000,
     "num_timesteps": 1000000
   },
diff --git a/config/config_A2C_Town02_mlp.json b/config/config_A2C_Town02_mlp.json
@@ -0,0 +1,24 @@
+{
+  "base":{
+    "seed": [0],
+    "eval_freq": 10000,
+    "num_timesteps": 1000000
+  },
+  "agent":{
+    "name": "A2C",
+    "discount": 0.99,
+    "alpha": [8e-05],
+    "buffer_size": 64
+  },
+  "environment":{
+    "map_name": "Town02",
+    "data_dir": "carla_data",
+    "start_state": 16,
+    "goal_states": [89,90],
+    "crosswalk_states": [96],
+    "r_base": 1, 
+    "r_loopback": 0
+  },
+  "policy": "MlpPolicy",
+  "save_dir": "Baselines"
+} 
diff --git a/config/config_A2C_Town03_mlp.json b/config/config_A2C_Town03_mlp.json
@@ -0,0 +1,24 @@
+{
+  "base":{
+    "seed": [0],
+    "eval_freq": 10000,
+    "num_timesteps": 2000000
+  },
+  "agent":{
+    "name": "A2C",
+    "discount": 0.99,
+    "alpha": [1e-04],
+    "buffer_size": 64
+  },
+  "environment":{
+    "map_name": "Town03",
+    "data_dir": "carla_data",
+    "start_state": 546,
+    "goal_states": [641, 642],
+    "crosswalk_states": [585],
+    "r_base": 1, 
+    "r_loopback": 0
+  },
+  "policy": "MlpPolicy",
+  "save_dir": "Baselines"
+} 
diff --git a/config/config_DQN_Town01_mlp.json b/config/config_DQN_Town01_mlp.json
@@ -0,0 +1,25 @@
+{
+  "base":{
+    "seed": [0],
+    "eval_freq": 10000,
+    "num_timesteps": 1000000
+  },
+  "agent":{
+    "name": "DQN",
+    "discount": 0.99,
+    "alpha": [1e-04],
+    "buffer_size": 2048,
+    "batch_size": 64
+  },
+  "environment":{
+    "map_name": "Town01",
+    "data_dir": "carla_data",
+    "start_state": 209,
+    "goal_states": [112, 113],
+    "crosswalk_states": [261],
+    "r_base": 1,
+    "r_loopback": 0
+  },
+  "policy": "MlpPolicy",
+  "save_dir": "Baselines"
+}
diff --git a/config/config_DQN_Town02_mlp.json b/config/config_DQN_Town02_mlp.json
@@ -0,0 +1,25 @@
+{
+  "base":{
+    "seed": [0],
+    "eval_freq": 10000,
+    "num_timesteps": 1000000
+  },
+  "agent":{
+    "name": "DQN",
+    "discount": 0.99,
+    "alpha": [5e-06],
+    "buffer_size": 2048,
+    "batch_size": 64
+  },
+  "environment":{
+    "map_name": "Town02",
+    "data_dir": "carla_data",
+    "start_state": 16,
+    "goal_states": [89,90],
+    "crosswalk_states": [96],
+    "r_base": 1,
+    "r_loopback": 0
+  },
+  "policy": "MlpPolicy",
+  "save_dir": "Baselines"
+}
diff --git a/config/config_DQN_Town03_mlp.json b/config/config_DQN_Town03_mlp.json
@@ -0,0 +1,25 @@
+{
+  "base":{
+    "seed": [0],
+    "eval_freq": 10000,
+    "num_timesteps": 2000000
+  },
+  "agent":{
+    "name": "DQN",
+    "discount": 0.99,
+    "alpha": [3e-04],
+    "buffer_size": 2048,
+    "batch_size": 64
+  },
+  "environment":{
+    "map_name": "Town03",
+    "data_dir": "carla_data",
+    "start_state": 546,
+    "goal_states": [641, 642],
+    "crosswalk_states": [585],
+    "r_base": 1,
+    "r_loopback": 0
+  },
+  "policy": "MlpPolicy",
+  "save_dir": "Baselines"
+}
diff --git a/config/config_PPO_Town01_mlp.json b/config/config_PPO_Town01_mlp.json
@@ -1,6 +1,6 @@
 {
   "base":{
-    "seed": [0,1,2,3,4],
+    "seed": [0],
     "eval_freq": 10000,
     "num_timesteps": 1000000
   },
diff --git a/config/config_PPO_Town02_mlp.json b/config/config_PPO_Town02_mlp.json
@@ -0,0 +1,26 @@
+{
+  "base":{
+    "seed": [0],
+    "eval_freq": 10000,
+    "num_timesteps": 1000000
+  },
+  "agent":{
+    "name": "PPO",
+    "discount": 0.99,
+    "alpha": [3e-04],
+    "buffer_size": 2048,
+    "batch_size": 64,
+    "n_epochs":1
+  },
+  "environment":{
+    "map_name": "Town02",
+    "data_dir": "carla_data",
+    "start_state": 16,
+    "goal_states": [89, 90],
+    "crosswalk_states": [96],
+    "r_base": 1,
+    "r_loopback": 0
+  },
+  "policy": "MlpPolicy",
+  "save_dir": "Baselines"
+}
diff --git a/config/config_PPO_Town03_mlp.json b/config/config_PPO_Town03_mlp.json
@@ -0,0 +1,26 @@
+{
+  "base":{
+    "seed": [0],
+    "eval_freq": 10000,
+    "num_timesteps": 2000000
+  },
+  "agent":{
+    "name": "PPO",
+    "discount": 0.99,
+    "alpha": [2e-04],
+    "buffer_size": 2048,
+    "batch_size": 64,
+    "n_epochs":1
+  },
+  "environment":{
+    "map_name": "Town03",
+    "data_dir": "carla_data",
+    "start_state": 546,
+    "goal_states": [641, 642],
+    "crosswalk_states": [585],
+    "r_base": 1,
+    "r_loopback": 0
+  },
+  "policy": "MlpPolicy",
+  "save_dir": "Baselines"
+}
diff --git a/config/config_QRDQN_Town01_mlp.json b/config/config_QRDQN_Town01_mlp.json
@@ -0,0 +1,30 @@
+{
+  "base":{
+    "seed": [0],
+    "eval_freq": 10000,
+    "num_timesteps": 2000000
+  },
+  "agent":{
+    "name": "QRDQN",
+    "discount": 0.99,
+    "alpha": [6e-04],
+    "buffer_size": 2048,
+    "batch_size": 64,
+    "n_quantiles": 4,
+    "epsilon": 0.1,
+    "eps_fraction": 0.02,
+    "eval_policy": "Greedy"
+  },
+  "environment":{
+    "map_name": "Town01",
+    "data_dir": "carla_data",
+    "start_state": 209,
+    "goal_states": [112, 113],
+    "crosswalk_states": [261],
+    "r_base": 1,
+    "r_loopback": 0
+  },
+  "policy": "MlpPolicy",
+  "save_dir": "Baselines"
+}
+
diff --git a/config/config_QRDQN_Town01_robust_rl_ssd.json b/config/config_QRDQN_Town01_robust_rl_ssd.json
@@ -0,0 +1,30 @@
+{
+  "base":{
+    "seed": [0],
+    "eval_freq": 10000,
+    "num_timesteps": 1000000
+  },
+  "agent":{
+    "name": "QRDQN",
+    "discount": 0.99,
+    "alpha": [5e-04],
+    "buffer_size": 2048,
+    "batch_size": 64,
+    "n_quantiles": 4,
+    "epsilon": 0.1,
+    "eps_fraction": 0.02,
+    "eval_policy": "SSD"
+  },
+  "environment":{
+    "map_name": "Town01",
+    "data_dir": "carla_data",
+    "start_state": 209,
+    "goal_states": [112, 113],
+    "crosswalk_states": [261],
+    "r_base": 3,
+    "r_loopback": 18
+  },
+  "policy": "CnnPolicy",
+  "save_dir": "Baselines"
+}
+
diff --git a/config/config_QRDQN_Town01_robust_rl_thres_ssd.json b/config/config_QRDQN_Town01_robust_rl_thres_ssd.json
@@ -0,0 +1,30 @@
+{
+  "base":{
+    "seed": [0],
+    "eval_freq": 10000,
+    "num_timesteps": 1000000
+  },
+  "agent":{
+    "name": "QRDQN",
+    "discount": 0.99,
+    "alpha": [5e-04],
+    "buffer_size": 2048,
+    "batch_size": 64,
+    "n_quantiles": 4,
+    "epsilon": 0.1,
+    "eps_fraction": 0.02,
+    "eval_policy": "Thresholded_SSD"
+  },
+  "environment":{
+    "map_name": "Town01",
+    "data_dir": "carla_data",
+    "start_state": 209,
+    "goal_states": [112, 113],
+    "crosswalk_states": [261],
+    "r_base": 3,
+    "r_loopback": 18
+  },
+  "policy": "CnnPolicy",
+  "save_dir": "Baselines"
+}
+
diff --git a/config/config_QRDQN_Town02_mlp.json b/config/config_QRDQN_Town02_mlp.json
@@ -0,0 +1,30 @@
+{
+  "base":{
+    "seed": [0],
+    "eval_freq": 10000,
+    "num_timesteps": 2000000
+  },
+  "agent":{
+    "name": "QRDQN",
+    "discount": 0.99,
+    "alpha": [6e-05],
+    "buffer_size": 2048,
+    "batch_size": 64,
+    "n_quantiles": 4,
+    "epsilon": 0.1,
+    "eps_fraction": 0.02,
+    "eval_policy": "Greedy"
+  },
+  "environment":{
+    "map_name": "Town02",
+    "data_dir": "carla_data",
+    "start_state": 16,
+    "goal_states": [89, 90],
+    "crosswalk_states": [96],
+    "r_base": 1,
+    "r_loopback": 0
+  },
+  "policy": "MlpPolicy",
+  "save_dir": "Baselines"
+}
+
diff --git a/config/config_QRDQN_Town02_robust_rl_greedy.json b/config/config_QRDQN_Town02_robust_rl_greedy.json
@@ -25,6 +25,6 @@
     "r_loopback": 18
   },
   "policy": "CnnPolicy",
-  "save_dir": "Baselines"
+  "save_dir": "/home/rfal/Stochastic_Road_Network/Baselines"
 }
 
diff --git a/config/config_QRDQN_Town02_robust_rl_ssd.json b/config/config_QRDQN_Town02_robust_rl_ssd.json
@@ -25,6 +25,6 @@
     "r_loopback": 18
   },
   "policy": "CnnPolicy",
-  "save_dir": "Baselines"
+  "save_dir": "/home/rfal/Stochastic_Road_Network/Baselines"
 }
 
diff --git a/config/config_QRDQN_Town02_robust_rl_thres_ssd.json b/config/config_QRDQN_Town02_robust_rl_thres_ssd.json
@@ -26,6 +26,6 @@
     "r_loopback": 18
   },
   "policy": "CnnPolicy",
-  "save_dir": "Baselines"
+  "save_dir": "/home/rfal/Stochastic_Road_Network/Baselines"
 }
 
diff --git a/config/config_QRDQN_Town03_mlp.json b/config/config_QRDQN_Town03_mlp.json
@@ -0,0 +1,30 @@
+{
+  "base":{
+    "seed": [0],
+    "eval_freq": 10000,
+    "num_timesteps": 2000000
+  },
+  "agent":{
+    "name": "QRDQN",
+    "discount": 0.99,
+    "alpha": [5e-05],
+    "buffer_size": 2048,
+    "batch_size": 64,
+    "n_quantiles": 4,
+    "epsilon": 0.1,
+    "eps_fraction": 0.02,
+    "eval_policy": "Greedy"
+  },
+  "environment":{
+    "map_name": "Town03",
+    "data_dir": "carla_data",
+    "start_state": 546,
+    "goal_states": [641, 642],
+    "crosswalk_states": [585],
+    "r_base": 1,
+    "r_loopback": 0
+  },
+  "policy": "MlpPolicy",
+  "save_dir": "Baselines"
+}
+

Original file line number	Diff line number	Diff line change
`@@ -1,6 +1,6 @@`
`1`	`1`	`{`
`2`	`2`	`"base":{`
`3`		`- "seed": [0,1,2,3,4],`
	`3`	`+ "seed": [0],`
`4`	`4`	`"eval_freq": 10000,`
`5`	`5`	`"num_timesteps": 1000000`
`6`	`6`	`},`