[pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

Author: pre-commit-ci[bot]

auto_round/experimental/transform/apply.py

@@ -56,21 +56,14 @@ def _apply_to_module(
     if config.location == "input":
         from .triton.mxfp4 import mxfp4_forward_kernel_wrapper
 
-        transform = build_transform(
-            **config.dict(),
-            device="cpu",
-            precision=module.dtype,
-            location="input"
-        )
+        transform = build_transform(**config.dict(), device="cpu", precision=module.dtype, location="input")
 
         def input_hook(_, args):
             input = args[0]
             # transform(input)
             orig_shape = input.shape
             x_flat = input.contiguous().flatten(end_dim=-2)
-            qdq_input, _ = mxfp4_forward_kernel_wrapper(
-                x_flat, transform.weight
-            )
+            qdq_input, _ = mxfp4_forward_kernel_wrapper(x_flat, transform.weight)
             return qdq_input.reshape(orig_shape)
 
         # for fused transform + quantization kernel

auto_round/experimental/transform/transforms.py

@@ -72,13 +72,17 @@ def forward(self, x: torch.Tensor):
         ori_shape = x.shape
         x = x.view(-1, self.size)
         return (
-            apply_transform_weight(
-                self.weight.to(device=x.device),
-                x.to(dtype=self.weight.dtype),
-                self.location,
-                self.module_type,
+            (
+                apply_transform_weight(
+                    self.weight.to(device=x.device),
+                    x.to(dtype=self.weight.dtype),
+                    self.location,
+                    self.module_type,
+                )
             )
-        ).to(x.dtype).view(ori_shape)
+            .to(x.dtype)
+            .view(ori_shape)
+        )
 
 
 TRANSFORMS = {
@@ -90,4 +94,3 @@ def forward(self, x: torch.Tensor):
 def build_transform(transform_type: str, **transform_kwargs):
     transform = TRANSFORMS[transform_type]
     return transform(**filter_kwarg_dict(transform.__init__, transform_kwargs))
-

auto_round/experimental/transform/utils/__init__.py

@@ -0,0 +1,2 @@
+# # Copyright (C) 2026 Intel Corporation
+# # SPDX-License-Identifier: Apache-2.0

auto_round/experimental/transform/utils/hadamard.py

@@ -1,11 +1,12 @@
+# # Copyright (C) 2026 Intel Corporation
+# # SPDX-License-Identifier: Apache-2.0
 
 import math
 from pathlib import Path
 
 import torch
 from safetensors import safe_open
 
-
 REPO_PATH = Path(__file__).parent / "hadamards.safetensors"
 
 
@@ -25,7 +26,7 @@ def deterministic_hadamard_matrix(
     Construct an n-by-n Hadamard matrix, using Sylvester's construction.
     `n` must be a power of 2.
 
-    Adapated from https://github.com/scipy/scipy/blob/v1.15.2/scipy/linalg/_special_matrices.py  # noqa: E501
+    Adapted from https://github.com/scipy/scipy/blob/v1.15.2/scipy/linalg/_special_matrices.py  # noqa: E501
 
     :param size: order of the matrix, must be a power of 2
     :param dtype: data type of matrix
@@ -59,7 +60,7 @@ def random_hadamard_matrix(
     `deterministic_hadamard_matrix` in that this function supports non powers of 2
     and randomization using a seeded generator
 
-    Adapated from https://github.com/facebookresearch/SpinQuant/blob/main/utils/hadamard_utils.py  # noqa: E501
+    Adapted from https://github.com/facebookresearch/SpinQuant/blob/main/utils/hadamard_utils.py  # noqa: E501
     Known matrices were retrieved from N. J. A. Sloane's Library of Hadamard Matrices http://www.neilsloane.com/hadamard/  # noqa: E501
 
     :param size: The dimension of the hamadard matrix

auto_round/experimental/transform/utils/matrix.py

@@ -1,11 +1,14 @@
-import torch
+# # Copyright (C) 2026 Intel Corporation
+# # SPDX-License-Identifier: Apache-2.0
 
+import torch
 
 __all__ = ["apply_transform_weight"]
 
 # note that apply_transform_weight reuses some code from
 # https://github.com/vllm-project/compressed-tensors/blob/main/src/compressed_tensors/transform/utils/matrix.py
 
+
 def apply_transform_weight(
     transform_weight: torch.Tensor,
     value: torch.Tensor,
@@ -75,7 +78,7 @@ def _multihead_matmul(A: torch.Tensor, B: torch.Tensor) -> torch.Tensor:
          [  B  ]
          [    B]]
 
-    This function will error out if the shapes are not evenly divisble
+    This function will error out if the shapes are not evenly divisible
 
     :param A: left-hand tensor
     :param B: right-hand tensor