Merge branch 'main' into patch2478

Author: svekars

.pyspelling.yml

@@ -19,7 +19,7 @@ matrix:
         - open: '\.\.\s+(figure|literalinclude|math|image|grid)::'
           close: '\n'
         # Exclude roles:
-        - open: ':(?:(class|py:mod|mod|func)):`'
+        - open: ':(?:(class|py:mod|mod|func|meth|obj)):`'
           content: '[^`]*'
           close: '`'
         # Exclude reStructuredText hyperlinks
@@ -70,7 +70,7 @@ matrix:
       - open: ':figure:.*'
         close: '\n'
       # Ignore reStructuredText roles
-      - open: ':(?:(class|file|func|math|ref|octicon)):`'
+      - open: ':(?:(class|file|func|math|ref|octicon|meth|obj)):`'
         content: '[^`]*'
         close: '`'
       - open: ':width:'

_static/img/pendulum.gif

@@ -206,14 +206,15 @@ Note: this code is taken from
 
         # Fuse Conv+BN and Conv+BN+Relu modules prior to quantization 
         # This operation does not change the numerics 
-        def fuse_model(self): 
+        def fuse_model(self, is_qat=False): 
+            fuse_modules = torch.ao.quantization.fuse_modules_qat if is_qat else torch.ao.quantization.fuse_modules
             for m in self.modules():  
                 if type(m) == ConvBNReLU: 
-                    torch.ao.quantization.fuse_modules(m, ['0', '1', '2'], inplace=True)
+                    fuse_modules(m, ['0', '1', '2'], inplace=True)
                 if type(m) == InvertedResidual: 
                     for idx in range(len(m.conv)):  
                         if type(m.conv[idx]) == nn.Conv2d:  
-                            torch.ao.quantization.fuse_modules(m.conv, [str(idx), str(idx + 1)], inplace=True)
+                            fuse_modules(m.conv, [str(idx), str(idx + 1)], inplace=True)
 
 2. Helper functions 
 ------------------- 
@@ -533,7 +534,7 @@ We fuse modules as before
 .. code:: python
 
     qat_model = load_model(saved_model_dir + float_model_file)  
-    qat_model.fuse_model()  
+    qat_model.fuse_model(is_qat=True)  
 
     optimizer = torch.optim.SGD(qat_model.parameters(), lr = 0.0001) 
     # The old 'fbgemm' is still available but 'x86' is the recommended default. 

_static/img/rollout_recurrent.png

@@ -62,6 +62,7 @@ Colab
 Conv
 ConvNet
 ConvNets
+customizable
 DCGAN
 DCGANs
 DDP

advanced_source/pendulum.py

@@ -312,14 +312,26 @@ What's new in PyTorch tutorials?
    :link: intermediate/mario_rl_tutorial.html
    :tags: Reinforcement-Learning
 
+.. customcarditem::
+   :header: Recurrent DQN
+   :card_description: Use TorchRL to train recurrent policies
+   :image: _static/img/rollout_recurrent.png
+   :link: intermediate/dqn_with_rnn_tutorial.html
+   :tags: Reinforcement-Learning
+
 .. customcarditem::
    :header: Code a DDPG Loss
    :card_description: Use TorchRL to code a DDPG Loss
    :image: _static/img/half_cheetah.gif
    :link: advanced/coding_ddpg.html
    :tags: Reinforcement-Learning
 
-
+.. customcarditem::
+   :header: Writing your environment and transforms
+   :card_description: Use TorchRL to code a Pendulum
+   :image: _static/img/pendulum.gif
+   :link: advanced/pendulum.html
+   :tags: Reinforcement-Learning
 
 .. Deploying PyTorch Models in Production
 
@@ -951,6 +963,7 @@ Additional Resources
    intermediate/reinforcement_q_learning
    intermediate/reinforcement_ppo
    intermediate/mario_rl_tutorial
+   advanced/pendulum
 
 .. toctree::
    :maxdepth: 2