Disable logging for quantization (num_bits etc) by default

Author: natuan

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

@@ -103,6 +103,7 @@ class QuantizationModifier(ScheduledModifier):
     |       freeze_bn_stats_epoch: 3.0
     |       model_fuse_fn_name: 'fuse_module'
     |       strict: True
+    |       verbose: False
 
     :param start_epoch: The epoch to start the modifier at
     :param scheme: Default QuantizationScheme to use when enabling quantization
@@ -133,6 +134,8 @@ class QuantizationModifier(ScheduledModifier):
         scheme_overrides or ignore are not found in a given module. Default True
     :param end_epoch: Disabled, setting to anything other than -1 will raise an
         exception. For compatibility with YAML serialization only.
+    :param verbose: if True, will log detailed information such as number of bits, batch
+        norm freezing etc. Default to False
     """
 
     def __init__(
@@ -148,6 +151,7 @@ def __init__(
         num_calibration_steps: Optional[int] = None,
         strict: bool = True,
         end_epoch: float = -1.0,
+        verbose: bool = False,
     ):
         raise_if_torch_quantization_not_available()
         if end_epoch != -1:
@@ -178,7 +182,7 @@ def __init__(
             self._model_fuse_fn_name = None
 
         self._strict = strict
-
+        self._verbose = verbose
         self._qat_enabled = False
         self._quantization_observer_disabled = False
         self._bn_stats_frozen = False
@@ -348,6 +352,22 @@ def strict(self, value: bool):
         """
         self._strict = value
 
+    @ModifierProp()
+    def verbose(self) -> bool:
+        """
+        :return: if True, will log detailed information such as number of bits, batch
+            norm freezing etc
+        """
+        return self._verbose
+
+    @strict.setter
+    def verbose(self, value: bool):
+        """
+        :params value: if True, will log detailed information such as number of bits,
+            batch norm freezing etc.
+        """
+        self._verbose = value
+
     def initialize(
         self,
         module: Module,
@@ -455,7 +475,8 @@ def _check_quantization_update(
             module.apply(freeze_bn_stats)
             self._bn_stats_frozen = True
 
-        self._log_quantization(module, epoch, steps_per_epoch)
+        if self._verbose:
+            self._log_quantization(module, epoch, steps_per_epoch)
 
     def _disable_quantization_observer_update_ready(self, epoch: float) -> bool:
         return (