Created using Colaboratory

zhazhax · Mar 1, 2021 · 89be043 · 89be043
1 parent afd6a02
commit 89be043
Showing 1 changed file with 7 additions and 7 deletions.
diff --git a/Reacher_Example.ipynb b/Reacher_Example.ipynb
@@ -43,7 +43,7 @@
         "id": "FGXyBBvL0dR2"
       },
       "source": [
-        "# **Part 1: Testing Task Definition**\r\n",
+        "# **Part 1: Testing Task Description**\r\n",
         "\r\n",
         "Reacher-v2 is a classic task in robotics since it performs one of the most fundamental skills: reaching. The goal of the [Reacher-v2](https://gym.openai.com/envs/Reacher-v2/) task is straightforward that we want to make a 2D robot arm reach a randomly located target point in the plane."
       ]
@@ -87,7 +87,7 @@
         "id": "DbamGVHC3AeW"
       },
       "source": [
-        "# **Part 2: Install all the packages through ElegantRL**"
+        "# **Part 2: Install ElegantRL**"
       ]
     },
     {
@@ -147,7 +147,7 @@
         "id": "UVdmpnK_3Zcn"
       },
       "source": [
-        "# **Part 3: Import packages**\r\n",
+        "# **Part 3: Import Packages**\r\n",
         "\r\n",
         "\r\n",
         "*   **elegantrl**\r\n",
@@ -179,7 +179,7 @@
         "id": "3n8zcgcn14uq"
       },
       "source": [
-        "# **Part 4: Specify agent and environment**\r\n",
+        "# **Part 4: Specify Agent and Environment**\r\n",
         "\r\n",
         "*   **args.agentrl**: firstly chooses one DRL algorithm to use, and the user is able to choose any agent from agent.py\r\n",
         "*   **args.env**: creates and preprocesses the environment, and the user can either customize own environment or preprocess environments from OpenAI Gym and PyBullet Gym from env.py.\r\n",
@@ -228,7 +228,7 @@
         "id": "z1j5kLHF2dhJ"
       },
       "source": [
-        "# **Part 5: Train and evaluate the model**\r\n",
+        "# **Part 5: Train and Evaluate the Agent**\r\n",
         "\r\n",
         "> The training and evaluating processes are all finished inside function **train_and_evaluate__multiprocessing()**, and the only parameter for it is **args**. It includes the fundamental objects in DRL:\r\n",
         "\r\n",
@@ -300,7 +300,7 @@
         "id": "JPXOxLSqh5cP"
       },
       "source": [
-        "Reference of the result stats:\r\n",
+        "Understanding the above results::\r\n",
         "*   **Step**: the total training steps.\r\n",
         "*  **MaxR**: the maximum reward.\r\n",
         "*   **avgR**: the average of the rewards.\r\n",
@@ -315,7 +315,7 @@
         "id": "QgNM-ozQcMhv"
       },
       "source": [
-        "# **Part 6: Result**"
+        "# **Part 6: Testing Results**"
       ]
     },
     {