Merge pull request #7 from incognite-lab/PNPnew

Pn pnew

myGym/__init__.py

@@ -1,3 +1,3 @@
 # This file was generated by the setup.py and will be rewritten at the next build
-__version__ = "0.1.dev114+g126fcd1"
+__version__ = "2.1.dev5+geb0cf18"
 __name__ = "myGym"

myGym/configs/train_pnp.json → myGym/configs/speed_reach.json

@@ -10,41 +10,43 @@
     "visgym"    :0,
     #Robot
     "robot"         :"kuka",
-    "robot_action"  :"joints",
-    "robot_init"    :[0.0, 0.3, 0.9],
+    "robot_action"  :"step",
+    "robot_init"    :[0.0, 0.4, 0.4],
+    "max_velocity"  :10,
+    "max_force"     :300,
+    "action_repeat" :1,
     #Task
-    "task_type"             :"pnp",
-    "task_objects"          :[{"init":{"obj_name":"cube_holes","fixed":0,"rand_rot":0, "sampling_area":[0,0.4,0.5,0.6,0.1,0.1]},
-                           "goal":{"obj_name":"crate","fixed":1,"rand_rot":0, "sampling_area":[-0.4,0,0.5, 0.7, 0, 0]}}],
-    "color_dict"            : {"crate":[[0.2,0.5,0.2,1]], "spoon":[[0.3,0.3,0.3,1]]},
+    "task_type"             :"reach",
+    "task_objects"          :[{"init":{"obj_name":"null"},
+                               "goal":{"obj_name":"cube_holes","fixed":1,"rand_rot":0, "sampling_area":[0.1, -0.1, 0.6, 0.7, 0.2, 0.1]}}],
     "used_objects"          :{"num_range":[0,0], "obj_list":[]},
     // Observation
     // actual_state options (pick one): "endeff_xyz", "endeff_6D" (robot end effector), "obj_xyz", "obj_6D", "vae", "yolact", "voxel" or "dope"
     // goal_state options (pick one): "obj_xyz", "obj_6D", "vae", "yolact", "voxel" or "dope"
     // additional_obs options (pick none/one/more): "joints_xyz", "joints_angles", "endeff_xyz", "endeff_6D", "touch", "distractor"
-    "observation"           : {"actual_state":"obj_xyz", "goal_state":"obj_xyz", "additional_obs":["endeff_6D", "touch"]},
+    "observation"           : {"actual_state":"obj_xyz", "goal_state":"obj_xyz", "additional_obs":["endeff_6D"]},
     #Distractor
     "distractors"           : {"list":null, "moveable":1, "constant_speed":0, "movement_dims":3,
-                                       "movement_endpoints":[-0.3, 0.3, 0.4, 0.7, 0.1, 0.3]},
+                                           "movement_endpoints":[-0.3, 0.3, 0.4, 0.7, 0.1, 0.3]},
     #Reward
-    "reward"        :"pnp",
+    "reward"        :"distance",
     "distance_type" :"euclidean",
     "vae_path"      :null,
     "yolact_path"   :null,
     "yolact_config" :null,
     #Train
     "train_framework"   :"tensorflow",
     "algo"              :"ppo2",
-    "num_networks"      :1,
-    "max_episode_steps" :512,
-    "algo_steps"        :512,
-    "steps"             :2000000,
+    "num_networks" : 1,
+    "max_episode_steps" :256,
+    "algo_steps"        :256,
+    "steps"             :20000,
     "pretrained_model"  :null,
     "multiprocessing"   :false,
     #Evaluation
-    "eval_freq"     :10000,
+    "eval_freq"     :21000,
     "eval_episodes" :10,
     #Saving and logging
-    "logdir"    :"trained_models/pnp",
+    "logdir"    :"trained_models/check",
     "record"    :0
-    }
+    }

myGym/configs/testicub.json

@@ -0,0 +1,52 @@
+{
+#Environment
+"env_name" :"Gym-v0",
+"workspace":"baskets",
+"engine"   :"pybullet",
+"render"   :"opengl",
+"camera"   :0,
+"gui"      :1,
+"visualize":0,
+"visgym"   :0,
+#Robot
+"robot"       :"icub",
+"robot_action":"joints",
+"robot_init"  :[0.0, 0.5, 0.2],
+"max_velocity"  :10,
+"max_force"     :70,
+"action_repeat" :1,
+
+#Task
+"task_type"           :"poke",
+"task_objects"          :[{"init":{"obj_name":"cube_poke","fixed":0,"rand_rot":0, "sampling_area":[0, 0, 0.6, 0.6, 0.1, 0.1]},
+                           "goal":{"obj_name":"target","fixed":1,"rand_rot":0, "sampling_area":[-0.3, 0.3, 0.8, 0.9, 0.05, 0.05]}}],
+"used_objects"          :{"num_range":[0,0], "obj_list":[]},
+// Observation
+// actual_state options (pick one): "endeff_xyz", "endeff_6D" (robot end effector), "obj_xyz", "obj_6D", "vae", "yolact", "voxel" or "dope"
+// goal_state options (pick one): "obj_xyz", "obj_6D", "vae", "yolact", "voxel" or "dope"
+// additional_obs options (pick none/one/more): "joints_xyz", "joints_angles", "endeff_xyz", "endeff_6D", "touch", "distractor"
+"observation"           : {"actual_state":"obj_xyz", "goal_state":"obj_xyz", "additional_obs":["endeff_xyz","touch","joints_angles"]},
+"distractors"           : {"list":null, "moveable":1, "constant_speed":1, "movement_dims":1,
+                                       "movement_endpoints":[-0.3, 0.3]},
+#Reward
+"reward"       :"poke",
+"distance_type":"euclidean",
+"vae_path"     :null,
+"yolact_path"  :null,
+"yolact_config":null,
+#Train
+"train_framework"  :"tensorflow",
+"algo"             :"ppo2",
+"num_networks"     : 1,
+"max_episode_steps":9024,
+"algo_steps"       :512,
+"steps"            :300000,
+"pretrained_model" :null,
+"multiprocessing"  :false,
+#Evaluation
+"eval_freq"    :10000,
+"eval_episodes":20,
+#Saving and logging
+"logdir": "trained_models/poke/",
+"record":0
+}

myGym/configs/multimodular_pnp.json → myGym/configs/train_pnp_3n.json

@@ -5,24 +5,27 @@
     "engine"    :"pybullet",
     "render"    :"opengl",
     "camera"    :0,
-    "gui"       :1,
+    "gui"       :0,
     "visualize" :0,
     "visgym"    :0,
     #Robot
-    "robot"         :"kuka",
-    "robot_action"  :"joints_gripper",
-    "robot_init"    :[0.0, 0.3, 0.9],
+    "robot"         :"panda1",
+    "robot_action"  :"step",
+    "robot_init"    :[-0.4, 0.4, 0.4],
+    "max_velocity"  :1,
+    "max_force"     :70,
+    "action_repeat" :1,
     #Task
     "task_type"             :"pnp",
-    "task_objects"          :[{"init":{"obj_name":"cube_holes","fixed":0,"rand_rot":0, "sampling_area":[0,0.4,0.5,0.6,0.1,0.1]},
-                           "goal":{"obj_name":"crate","fixed":1,"rand_rot":0, "sampling_area":[-0.4,0,0.5, 0.7, 0, 0]}}],
-    "color_dict"            : {"crate":[[0.2,0.5,0.2,1]], "spoon":[[0.3,0.3,0.3,1]]},
+    "task_objects"          :[{"init":{"obj_name":"cube_holes","fixed":0,"rand_rot":0, "sampling_area":[-0.38,-0.38,0.42,0.42,0.1,0.1]},
+                           "goal":{"obj_name":"cube_target","fixed":1,"rand_rot":0, "sampling_area":[0.3,-0.2,0.4,0.7,0.15,0.15]}}],
+    "color_dict"            : {"cube_holes":[[0.2,0.5,0.2,1]], "target":[[0.3,0.3,0.3,1]]},
     "used_objects"          :{"num_range":[0,0], "obj_list":[]},
     // Observation
     // actual_state options (pick one): "endeff_xyz", "endeff_6D" (robot end effector), "obj_xyz", "obj_6D", "vae", "yolact", "voxel" or "dope"
     // goal_state options (pick one): "obj_xyz", "obj_6D", "vae", "yolact", "voxel" or "dope"
     // additional_obs options (pick none/one/more): "joints_xyz", "joints_angles", "endeff_xyz", "endeff_6D", "touch", "distractor"
-    "observation"           : {"actual_state":"obj_xyz", "goal_state":"obj_xyz", "additional_obs":["endeff_6D", "touch"]},
+    "observation"           : {"actual_state":"obj_xyz", "goal_state":"obj_xyz", "additional_obs":["endeff_xyz","touch"]},
     #Distractor
     "distractors"           : {"list":null, "moveable":1, "constant_speed":0, "movement_dims":3,
                                        "movement_endpoints":[-0.3, 0.3, 0.4, 0.7, 0.1, 0.3]},
@@ -35,17 +38,16 @@
     #Train
     "train_framework"   :"tensorflow",
     "algo"              :"multi",
-    "num_networks"      : 3,
-    "network_swtcher"   : "gt",
+    "num_networks"      :3,
     "max_episode_steps" :512,
     "algo_steps"        :512,
-    "steps"             :100000,
+    "steps"             :5100000,
     "pretrained_model"  :null,
     "multiprocessing"   :false,
     #Evaluation
-    "eval_freq"     :10000,
-    "eval_episodes" :10,
+    "eval_freq"     :500000,
+    "eval_episodes" :100,
     #Saving and logging
-    "logdir"    :"trained_models/dual/PNP",
+    "logdir"    :"trained_models/pnp2nmagnet",
     "record"    :0
     }