fix bug

internlm/solver/optimizer/hybrid_zero_optim.py

@@ -817,6 +817,8 @@ class HybridZeroOptimizer(BaseOptimizer):
             self.optim.step()
             # release the fp32 grad
             for group_id in range(self.num_param_groups):
+                if not self.param_group_has_params[group_id]:
+                    continue
                 release_param_grad(self._fp32_orig_param_groups_of_current_rank[group_id])
 
             if self._enable_memory_balance and self._memory_balance_role == 1: