add self.reset() at the bottom of __init__ in class VQBeTPolicy

lerobot/common/policies/vqbet/modeling_vqbet.py

@@ -61,9 +61,7 @@ class VQBeTPolicy(nn.Module, PyTorchModelHubMixin):
 
         self.vqbet = VQBeTModel(config)
 
-    def check_discretized(self):
-        return self.vqbet.action_head.vqvae_model.discretized
-
+        self.reset()
 
     def reset(self):
         """
@@ -75,6 +73,9 @@ class VQBeTPolicy(nn.Module, PyTorchModelHubMixin):
             "action": deque(maxlen=self.config.n_action_pred_chunk),
         }
 
+    def check_discretized(self):
+        return self.vqbet.action_head.vqvae_model.discretized
+
     @torch.no_grad
     def select_action(self, batch: dict[str, Tensor]) -> Tensor:
         """Select a single action given environment observations.