Comments

openvinotoolkit · Oct 11, 2024 · 6d2ecc0 · 6d2ecc0
1 parent 455a23d
commit 6d2ecc0
Show file tree

Hide file tree

Showing 4 changed files with 7,521 additions and 8,915 deletions.
diff --git a/nncf/experimental/torch/fx/transformations.py b/nncf/experimental/torch/fx/transformations.py
@@ -21,6 +21,7 @@
 import nncf
 from nncf.common.graph.graph import NNCFNode
 from nncf.common.graph.transformations.commands import TargetType
+from nncf.experimental.torch.fx.constant_folding import constant_fold
 from nncf.experimental.torch.fx.node_utils import get_graph_node_by_name
 from nncf.experimental.torch.fx.node_utils import get_tensor_constant_from_node
 from nncf.torch.graph.transformations.commands import PTTargetPoint
@@ -537,6 +538,7 @@ def apply_quantization_transformations(model: torch.fx.GraphModule) -> None:
     # to make it easier for algorithms to work
     # with the target graph BatchNorm operations
     # are being fused
+    constant_fold(model)
     fuse_conv_bn(model)
     separate_conv_and_bias(model)
     separate_linear_and_bias(model)

diff --git a/tests/post_training/pipelines/image_classification_torchvision.py b/tests/post_training/pipelines/image_classification_torchvision.py
@@ -19,7 +19,6 @@
 from torch._export import capture_pre_autograd_graph
 from torchvision import models
 
-from nncf.experimental.torch.fx.constant_folding import constant_fold
 from nncf.torch import disable_patching
 from tests.post_training.pipelines.base import PT_BACKENDS
 from tests.post_training.pipelines.base import BackendType
@@ -80,7 +79,6 @@ def prepare_model(self) -> None:
             with torch.no_grad():
                 with disable_patching():
                     self.model = self.model_params.export_fn(model, (self.dummy_tensor,))
-                    constant_fold(self.model)
 
         elif self.backend in PT_BACKENDS:
             self.model = model