Fix for legacy quantization modifier w/ DDP (#1490)

* Remove "module." added by DDP when identifying the submodules for quantization

* Style and quality fixes

---------

Co-authored-by: Benjamin Fineran <bfineran@users.noreply.github.com>

src/sparseml/pytorch/sparsification/quantization/legacy_modifier_quantization.py

@@ -504,6 +504,8 @@ def initialize(
         if self._submodules is not None:
             found_submodules = []
             for name, submodule in module.named_modules():
+                if name.startswith("module."):
+                    name = name[7:]
                 if name in self._submodules:
                     self._modules_to_quantize.append(_ModuleToQuantize(name, submodule))
                     found_submodules.append(name)