try 8x mem for cloud

Project/Assets/ML-Agents/Examples/Walker/Demos/ExpertWalkerDyn.demo.meta

 guid: 870f0916603374b429c92b4e74d912e1
 ScriptedImporter:
   fileIDToRecycleName:
-    11400000: Assets/Demonstrations/ExpertWalkerDyn.demo
+    11400002: Assets/ML-Agents/Examples/Walker/Demos/ExpertWalkerDyn.demo
   externalObjects: {}
   userData: ' (Unity.MLAgents.Demonstrations.DemonstrationSummary)'
   assetBundleName: 

Project/Assets/ML-Agents/Examples/Walker/Scenes/WalkerDynamic.unity

     - target: {fileID: 4712600297668500197, guid: f51e8260728fd4c8fa87bcda9d0e2027,
         type: 3}
       propertyPath: m_IsActive
-      value: 0
+      value: 1
       objectReference: {fileID: 0}
     - target: {fileID: 4715983678655513245, guid: f51e8260728fd4c8fa87bcda9d0e2027,
         type: 3}
     - target: {fileID: 4712600297668500197, guid: f51e8260728fd4c8fa87bcda9d0e2027,
         type: 3}
       propertyPath: m_IsActive
-      value: 0
+      value: 1
       objectReference: {fileID: 0}
     - target: {fileID: 4715983678655513245, guid: f51e8260728fd4c8fa87bcda9d0e2027,
         type: 3}
     - target: {fileID: 4712600297668500197, guid: f51e8260728fd4c8fa87bcda9d0e2027,
         type: 3}
       propertyPath: m_IsActive
-      value: 0
+      value: 1
       objectReference: {fileID: 0}
     - target: {fileID: 4715983678655513245, guid: f51e8260728fd4c8fa87bcda9d0e2027,
         type: 3}
     - target: {fileID: 4712600297668500197, guid: f51e8260728fd4c8fa87bcda9d0e2027,
         type: 3}
       propertyPath: m_IsActive
-      value: 0
+      value: 1
       objectReference: {fileID: 0}
     - target: {fileID: 4715983678655513245, guid: f51e8260728fd4c8fa87bcda9d0e2027,
         type: 3}
     - target: {fileID: 4712600297668500197, guid: f51e8260728fd4c8fa87bcda9d0e2027,
         type: 3}
       propertyPath: m_IsActive
-      value: 0
+      value: 1
       objectReference: {fileID: 0}
     - target: {fileID: 4715983678655513245, guid: f51e8260728fd4c8fa87bcda9d0e2027,
         type: 3}
     - target: {fileID: 4712600297668500197, guid: f51e8260728fd4c8fa87bcda9d0e2027,
         type: 3}
       propertyPath: m_IsActive
-      value: 0
+      value: 1
       objectReference: {fileID: 0}
     - target: {fileID: 4715983678655513245, guid: f51e8260728fd4c8fa87bcda9d0e2027,
         type: 3}
     - target: {fileID: 4712600297668500197, guid: f51e8260728fd4c8fa87bcda9d0e2027,
         type: 3}
       propertyPath: m_IsActive
-      value: 0
+      value: 1
       objectReference: {fileID: 0}
     - target: {fileID: 4715983678655513245, guid: f51e8260728fd4c8fa87bcda9d0e2027,
         type: 3}
     - target: {fileID: 4712600297668500197, guid: f51e8260728fd4c8fa87bcda9d0e2027,
         type: 3}
       propertyPath: m_IsActive
-      value: 0
+      value: 1
       objectReference: {fileID: 0}
     - target: {fileID: 4715983678655513245, guid: f51e8260728fd4c8fa87bcda9d0e2027,
         type: 3}
     - target: {fileID: 4712600297668500197, guid: f51e8260728fd4c8fa87bcda9d0e2027,
         type: 3}
       propertyPath: m_IsActive
-      value: 0
+      value: 1
       objectReference: {fileID: 0}
     - target: {fileID: 4715983678655513245, guid: f51e8260728fd4c8fa87bcda9d0e2027,
         type: 3}
     - target: {fileID: 4712600297668500197, guid: f51e8260728fd4c8fa87bcda9d0e2027,
         type: 3}
       propertyPath: m_IsActive
-      value: 0
+      value: 1
       objectReference: {fileID: 0}
     - target: {fileID: 4715983678655513245, guid: f51e8260728fd4c8fa87bcda9d0e2027,
         type: 3}
     - target: {fileID: 4712600297668500197, guid: f51e8260728fd4c8fa87bcda9d0e2027,
         type: 3}
       propertyPath: m_IsActive
-      value: 0
+      value: 1
       objectReference: {fileID: 0}
     - target: {fileID: 4715983678655513245, guid: f51e8260728fd4c8fa87bcda9d0e2027,
         type: 3}

Project/ProjectSettings/TimeManager.asset

   m_ObjectHideFlags: 0
   Fixed Timestep: 0.02
   Maximum Allowed Timestep: 0.33333334
-  m_TimeScale: 1
+  m_TimeScale: 20
   Maximum Particle Timestep: 0.03

config/ppo/WalkerDynamic.yaml

     trainer: ppo
     batch_size: 2048
     beta: 0.005
-    buffer_size: 20480
+    buffer_size: 163840 #20480
     epsilon: 0.2
     hidden_units: 512
     lambd: 0.95

config/trainer_config.yaml

             strength: 1.0
             gamma: 0.995
 
-WalkerDynamic:
-    normalize: true
-    # learning_rate_schedule: constant
-    num_epoch: 6
-    time_horizon: 2048
-    batch_size: 4096
-    buffer_size: 20480
-    max_steps: 2e7
-    # max_steps: 1e8
-    summary_freq: 30000
-    # learning_rate: 1.0e-3
-    num_layers: 3
-    hidden_units: 512
-    reward_signals:
-        extrinsic:
-            strength: 1.0
-            gamma: 0.995
-
+#     # learning_rate_schedule: constant
-#     time_horizon: 1000
-#     batch_size: 2048
-#     buffer_size: 20480
+#     time_horizon: 80
+#     batch_size: 512
+#     buffer_size: 5120
+#     # learning_rate: 1.0e-3
 #     num_layers: 3
 #     hidden_units: 512
 #     reward_signals:
+
+WalkerDynamic:
+    normalize: true
+    num_epoch: 3
+    time_horizon: 1000
+    batch_size: 2048
+    buffer_size: 20480
+    max_steps: 2e7
+    summary_freq: 30000
+    num_layers: 3
+    hidden_units: 512
+    reward_signals:
+        extrinsic:
+            strength: 1.0
+            gamma: 0.995
 
 WalkerStatic:
     normalize: true