style: silence F821 on calls to removed transform_with_custom_compression_ops

The 8 call sites are in @pytest.mark.skip'd tests for a removed API; add
# noqa: F821 so ruff passes until the obsolete tests are rewritten or removed.

Author: DawerG

tests/test_externalize.py

@@ -3065,7 +3065,7 @@ def forward(self, x: torch.Tensor) -> torch.Tensor:
     )
     quantizer = PostTrainingQuantizer(model, quantization_config)
     model = cast("nn.Module", quantizer.compress())
-    transform_with_custom_compression_ops(model)
+    transform_with_custom_compression_ops(model)  # noqa: F821
 
     sample = (torch.randn(2, 24),)
 
@@ -3153,7 +3153,7 @@ def forward(self, x: torch.Tensor) -> torch.Tensor:
     )
     quantizer = PostTrainingQuantizer(model, quantization_config)
     model = cast("nn.Module", quantizer.compress())
-    transform_with_custom_compression_ops(model)
+    transform_with_custom_compression_ops(model)  # noqa: F821
 
     sample = (torch.randn(2, 24),)
 
@@ -3217,7 +3217,7 @@ def forward(self, x: torch.Tensor, indices: torch.Tensor) -> torch.Tensor:
     )
     quantizer = PostTrainingQuantizer(model, quantization_config)
     model = cast("nn.Module", quantizer.compress())
-    transform_with_custom_compression_ops(model)
+    transform_with_custom_compression_ops(model)  # noqa: F821
 
     x = torch.randn(2, 1, 1, in_dim)
     indices = torch.tensor([[0, 2], [1, 3]], dtype=torch.int16)
@@ -3318,7 +3318,7 @@ def forward(self, x: torch.Tensor, indices: torch.Tensor) -> torch.Tensor:
     )
     quantizer = PostTrainingQuantizer(model, quantization_config)
     model = cast("nn.Module", quantizer.compress())
-    transform_with_custom_compression_ops(model)
+    transform_with_custom_compression_ops(model)  # noqa: F821
 
     x = torch.randn(2, 1, 1, in_dim)
     indices = torch.tensor([[0, 2], [1, 3]], dtype=torch.int16)
@@ -3390,7 +3390,7 @@ def forward(self, x: torch.Tensor) -> torch.Tensor:
     )
     quantizer = PostTrainingQuantizer(model, quantization_config)
     model = cast("nn.Module", quantizer.compress())
-    transform_with_custom_compression_ops(model)
+    transform_with_custom_compression_ops(model)  # noqa: F821
 
     sample = (torch.randn(1, 4, embed_dim),)
 
@@ -3484,7 +3484,7 @@ def forward(self, x: torch.Tensor) -> torch.Tensor:
     )
     quantizer = PostTrainingQuantizer(model, quantization_config)
     model = cast("nn.Module", quantizer.compress())
-    transform_with_custom_compression_ops(model)
+    transform_with_custom_compression_ops(model)  # noqa: F821
 
     sample = (torch.randn(1, 4, embed_dim),)
 
@@ -3553,7 +3553,7 @@ def forward(self, x: torch.Tensor, indices: torch.Tensor) -> torch.Tensor:
     )
     quantizer = PostTrainingQuantizer(model, quantization_config)
     model = cast("nn.Module", quantizer.compress())
-    transform_with_custom_compression_ops(model)
+    transform_with_custom_compression_ops(model)  # noqa: F821
 
     x = torch.randn(2, in_dim)
     indices = torch.tensor([[0, 2], [1, 3]], dtype=torch.int16)
@@ -3637,7 +3637,7 @@ def forward(self, x: torch.Tensor, indices: torch.Tensor) -> torch.Tensor:
     )
     quantizer = PostTrainingQuantizer(model, quantization_config)
     model = cast("nn.Module", quantizer.compress())
-    transform_with_custom_compression_ops(model)
+    transform_with_custom_compression_ops(model)  # noqa: F821
 
     x = torch.randn(2, in_dim)
     indices = torch.tensor([[0, 2], [1, 3]], dtype=torch.int16)