Add link to PR

Author: araffin

docs/modules/a2c.rst

@@ -93,7 +93,7 @@ Train a A2C agent on ``CartPole-v1`` using 4 environments.
 
 .. note::
 
-  **Using gSDE (Generalized State-Dependent Exploration) during inference:**
+  Using gSDE (Generalized State-Dependent Exploration) during inference (see `PR #1767 <https://github.com/DLR-RM/stable-baselines3/pull/1767>`_):
 
   When using A2C models trained with ``use_sde=True``, the automatic noise resetting that occurs during training (controlled by ``sde_sample_freq``) does not happen when using ``model.predict()`` for inference. This results in deterministic behavior even when ``deterministic=False``.
 

docs/modules/ppo.rst

@@ -107,7 +107,7 @@ Train a PPO agent on ``CartPole-v1`` using 4 environments.
 
 .. note::
 
-  **Using gSDE (Generalized State-Dependent Exploration) during inference:**
+  Using gSDE (Generalized State-Dependent Exploration) during inference (see `PR #1767 <https://github.com/DLR-RM/stable-baselines3/pull/1767>`_):
 
   When using PPO models trained with ``use_sde=True``, the automatic noise resetting that occurs during training (controlled by ``sde_sample_freq``) does not happen when using ``model.predict()`` for inference. This results in deterministic behavior even when ``deterministic=False``.
 

docs/modules/sac.rst

@@ -95,7 +95,7 @@ This example is only to demonstrate the use of the library and its functions, an
 
 .. note::
 
-  **Using gSDE (Generalized State-Dependent Exploration) during inference:**
+  Using gSDE (Generalized State-Dependent Exploration) during inference (see `PR #1767 <https://github.com/DLR-RM/stable-baselines3/pull/1767>`_):
 
   When using SAC models trained with ``use_sde=True``, the automatic noise resetting that occurs during training (controlled by ``sde_sample_freq``) does not happen when using ``model.predict()`` for inference. This results in deterministic behavior even when ``deterministic=False``.