Add policy gradient, reduce steps for runtime, fix PR comment

KIT-MRT · Dec 21, 2024 · 41cabce · 41cabce
1 parent 75136f4
commit 41cabce
Show file tree

Hide file tree

Showing 3 changed files with 16 additions and 7 deletions.
diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml
@@ -6,6 +6,9 @@ on:
   schedule:
     - cron:  "0 3 * * 1" # Run every Monday 3 am
 
+permissions:
+  pull-requests: write
+
 jobs:
   ci:
     runs-on: ubuntu-latest
@@ -62,10 +65,15 @@ jobs:
           publish_branch: gh-pages
           keep_files: true # Retain other content on gh-pages
 
-      - name: Add PR comment with preview link
+      - name: Add PR note
         if: github.event_name == 'pull_request'
-        run: |
-            echo "Deployed [https://${{ github.repository_owner }}.github.io/${{ github.event.repository.name }}/pr-${{ github.event.pull_request.number }}/](temporary docs)."
+        uses: marocchino/sticky-pull-request-comment@v2
+        with:
+          message: |
+            ### :books: Docs
+
+            Created [https://${{ github.repository_owner }}.github.io/${{ github.event.repository.name }}/pr-${{ github.event.pull_request.number }}/](temporary docs).
+          
 
 
 
diff --git a/mkdocs.yml b/mkdocs.yml
@@ -45,6 +45,7 @@ nav:
       - Decision Making: 
           - Value Iteration: notebooks/mdp_value_iteration.ipynb
           - Q-Learning: notebooks/mdp_q_learning.ipynb
+          - Policy Gradient: notebooks/mdp_policy_gradient.ipynb
   - API Documentation (partial): reference/
 
 plugins:

diff --git a/notebooks/mdp_policy_gradient.ipynb b/notebooks/mdp_policy_gradient.ipynb
@@ -62,7 +62,7 @@
     "model_checkpoints = policy_gradient(\n",
     "    mdp=grid_mdp,\n",
     "    policy=policy,\n",
-    "    iterations=100,\n",
+    "    iterations=10,\n",
     "    return_history=True,\n",
     ")"
    ]
@@ -107,7 +107,7 @@
     "    w = ipywidgets.interactive(plot_policy_step_grid_map, iteration=iteration_slider)\n",
     "    display(w)\n",
     "else:\n",
-    "    plot_policy_step_grid_map(100)"
+    "    plot_policy_step_grid_map(10)"
    ]
   },
   {
@@ -161,7 +161,7 @@
     "model_checkpoints = policy_gradient(\n",
     "    mdp=highway_mdp,\n",
     "    policy=policy,\n",
-    "    iterations=200,\n",
+    "    iterations=20,\n",
     "    return_history=True,\n",
     ")"
    ]
@@ -205,7 +205,7 @@
     "    w = ipywidgets.interactive(plot_policy_step_grid_map, iteration=iteration_slider)\n",
     "    display(w)\n",
     "else:\n",
-    "    plot_policy_step_grid_map(200)"
+    "    plot_policy_step_grid_map(20)"
    ]
   }
  ],