Update torchrl/objectives/value/advantages.py

Co-authored-by: Vincent Moens <vincentmoens@gmail.com>
pytorch · Nov 22, 2023 · 886b4e0 · 886b4e0
1 parent c23401a
commit 886b4e0
Showing 1 changed file with 1 addition and 0 deletions.
diff --git a/torchrl/objectives/value/advantages.py b/torchrl/objectives/value/advantages.py
@@ -1361,6 +1361,7 @@ class VTrace(ValueEstimatorBase):
         value_network (TensorDictModule): value operator used to retrieve the value estimates.
         actor_network (TensorDictModule): actor operator used to retrieve the log prob.
         rho_thresh (Union[float, Tensor]): rho clipping parameter for importance weights.
+            Defaults to ``1.0``.
         c_thresh (Union[float, Tensor]): c clipping parameter for importance weights.
         average_adv (bool): if ``True``, the resulting advantage values will be standardized.
             Default is ``False``.