minor

Signed-off-by: realAsma <akuriparambi@nvidia.com>

Author: realAsma

modelopt/torch/quantization/algorithms.py

@@ -961,6 +961,8 @@ def run_search_with_stats(self, max_weight_size, verbose=False):
         AutoQuantize uses Linear Programming Solver to find the optimal quantization configuration which
         minimizes the sum of per-layer auto_quantize scores while meeting the specified constraint.
         """
+        # TODO: Do this only for rank 0 in the respective pipeline group
+
         for lower_bound in [None, 0.99, 0.90]:
             # The LP solver for auto_quantize sometimes fails to find a solution if a lower bound is not
             # specified. I dont know why this happens.