Release 3 Cherry-pick bug-fixes and doc changes from master (#4102)

* [bug-fix] Fix regression in --initialize-from feature (#4086)

* Fixed text in GettingStarted page specifying the logdir for tensorboard. Before it was in a directory summaries which no longer existed. Results are now saved to the results dir. (#4085)

* [refactor] Remove nonfunctional `output_path` option from TrainerSettings (#4087)

* Reverting bug introduced in #4071 (#4101)

Co-authored-by: Scott <[email protected]>
Co-authored-by: Vincent-Pierre BERGES <[email protected]>

Author: 3 people

com.unity.ml-agents/Runtime/Academy.cs

@@ -366,8 +366,6 @@ void InitializeEnvironment()
                         port = port
                     }
                 );
-                Communicator.QuitCommandReceived += OnQuitCommandReceived;
-                Communicator.ResetCommandReceived += OnResetCommand;
             }
 
             if (Communicator != null)
@@ -400,6 +398,11 @@ void InitializeEnvironment()
                     Communicator = null;
                 }
             }
+            if (Communicator != null)
+            {
+                Communicator.QuitCommandReceived += OnQuitCommandReceived;
+                Communicator.ResetCommandReceived += OnResetCommand;
+            }
 
             // If a communicator is enabled/provided, then we assume we are in
             // training mode. In the absence of a communicator, we assume we are

config/imitation/CrawlerStatic.yaml

@@ -24,7 +24,6 @@ behaviors:
         use_actions: false
         use_vail: false
         demo_path: Project/Assets/ML-Agents/Examples/Crawler/Demos/ExpertCrawlerSta.demo
-    output_path: default
     keep_checkpoints: 5
     max_steps: 10000000
     time_horizon: 1000

config/imitation/FoodCollector.yaml

@@ -24,7 +24,6 @@ behaviors:
         use_actions: false
         use_vail: false
         demo_path: Project/Assets/ML-Agents/Examples/FoodCollector/Demos/ExpertFood.demo
-    output_path: default
     keep_checkpoints: 5
     max_steps: 2000000
     time_horizon: 64

config/imitation/Hallway.yaml

@@ -30,7 +30,6 @@ behaviors:
         use_actions: false
         use_vail: false
         demo_path: Project/Assets/ML-Agents/Examples/Hallway/Demos/ExpertHallway.demo
-    output_path: default
     keep_checkpoints: 5
     max_steps: 10000000
     time_horizon: 64

config/imitation/PushBlock.yaml

@@ -24,7 +24,6 @@ behaviors:
         use_actions: false
         use_vail: false
         demo_path: Project/Assets/ML-Agents/Examples/PushBlock/Demos/ExpertPush.demo
-    output_path: default
     keep_checkpoints: 5
     max_steps: 15000000
     time_horizon: 64

config/ppo/3DBall.yaml

@@ -19,7 +19,6 @@ behaviors:
       extrinsic:
         gamma: 0.99
         strength: 1.0
-    output_path: default
     keep_checkpoints: 5
     max_steps: 500000
     time_horizon: 1000

config/ppo/3DBallHard.yaml

@@ -19,7 +19,6 @@ behaviors:
       extrinsic:
         gamma: 0.995
         strength: 1.0
-    output_path: default
     keep_checkpoints: 5
     max_steps: 5000000
     time_horizon: 1000

config/ppo/3DBall_randomize.yaml

@@ -19,7 +19,6 @@ behaviors:
       extrinsic:
         gamma: 0.99
         strength: 1.0
-    output_path: default
     keep_checkpoints: 5
     max_steps: 500000
     time_horizon: 1000

config/ppo/Basic.yaml

@@ -19,7 +19,6 @@ behaviors:
       extrinsic:
         gamma: 0.9
         strength: 1.0
-    output_path: default
     keep_checkpoints: 5
     max_steps: 500000
     time_horizon: 3

config/ppo/Bouncer.yaml

@@ -19,7 +19,6 @@ behaviors:
       extrinsic:
         gamma: 0.99
         strength: 1.0
-    output_path: default
     keep_checkpoints: 5
     max_steps: 4000000
     time_horizon: 64