onnx 1.16.2__cp312-cp312-win_amd64.whl → 1.18.0__cp312-cp312-win_amd64.whl

onnx/reference/ops/op_greater_or_equal.py

@@ -1,7 +1,7 @@
 # Copyright (c) ONNX Project Contributors
 
 # SPDX-License-Identifier: Apache-2.0
-
+from __future__ import annotations
 
 import numpy as np
 

onnx/reference/ops/op_grid_sample.py

@@ -1,10 +1,9 @@
 # Copyright (c) ONNX Project Contributors
 
 # SPDX-License-Identifier: Apache-2.0
-
+from __future__ import annotations
 
 import numbers
-from typing import List
 
 import numpy as np
 
@@ -205,7 +204,7 @@ class GridSample(OpRun):
             return hi
         return val
 
-    def _pixel_at_ndarray(self, ndarray, x: List, border, padding_mode):  # type: ignore
+    def _pixel_at_ndarray(self, ndarray, x: list, border, padding_mode):  # type: ignore
         # boarder: [x_1_min, x_2_min, ..., x_1_max, x_2_max, ...]
         num_dims = ndarray.ndim
         assert num_dims == len(x) == int(len(border) / 2)

onnx/reference/ops/op_gru.py

@@ -1,7 +1,7 @@
 # Copyright (c) ONNX Project Contributors
 
 # SPDX-License-Identifier: Apache-2.0
-
+from __future__ import annotations
 
 import numpy as np
 
@@ -77,14 +77,14 @@ class CommonGRU(OpRun):
         B=None,
         sequence_lens=None,
         initial_h=None,
-        activation_alpha=None,
-        activation_beta=None,
-        activations=None,
-        clip=None,
-        direction=None,
+        activation_alpha=None,  # noqa: ARG002
+        activation_beta=None,  # noqa: ARG002
+        activations=None,  # noqa: ARG002
+        clip=None,  # noqa: ARG002
+        direction=None,  # noqa: ARG002
         hidden_size=None,
         layout=None,
-        linear_before_reset=None,
+        linear_before_reset=None,  # noqa: ARG002
     ):
         # TODO: support overridden attributes.
         num_directions = W.shape[0]

onnx/reference/ops/op_hamming_window.py

@@ -1,7 +1,7 @@
 # Copyright (c) ONNX Project Contributors
 
 # SPDX-License-Identifier: Apache-2.0
-
+from __future__ import annotations
 
 import numpy as np
 

onnx/reference/ops/op_hann_window.py

@@ -1,7 +1,7 @@
 # Copyright (c) ONNX Project Contributors
 
 # SPDX-License-Identifier: Apache-2.0
-
+from __future__ import annotations
 
 import numpy as np
 

onnx/reference/ops/op_hard_sigmoid.py

@@ -1,7 +1,7 @@
 # Copyright (c) ONNX Project Contributors
 
 # SPDX-License-Identifier: Apache-2.0
-
+from __future__ import annotations
 
 import numpy as np
 

onnx/reference/ops/op_hardmax.py

@@ -1,7 +1,7 @@
 # Copyright (c) ONNX Project Contributors
 
 # SPDX-License-Identifier: Apache-2.0
-
+from __future__ import annotations
 
 import numpy as np
 
@@ -14,6 +14,9 @@ class Hardmax(OpRunUnaryNum):
         x_argmax = np.argmax(x, axis=axis)  # type: ignore
         y = np.zeros_like(x)
         np.put_along_axis(
-            y, np.expand_dims(x_argmax, axis=axis), 1, axis=axis  # type: ignore
+            y,
+            np.expand_dims(x_argmax, axis=axis),
+            1,
+            axis=axis,  # type: ignore
         )
         return (y,)

onnx/reference/ops/op_identity.py

@@ -1,12 +1,12 @@
 # Copyright (c) ONNX Project Contributors
 
 # SPDX-License-Identifier: Apache-2.0
+from __future__ import annotations
 
+from onnx.reference.op_run import OpRun
 
-from onnx.reference.ops._op import OpRunUnaryNum
 
-
-class Identity(OpRunUnaryNum):
+class Identity(OpRun):
     def _run(self, a):  # type: ignore
         if a is None:
             return (None,)

onnx/reference/ops/op_if.py

@@ -4,10 +4,13 @@
 
 from __future__ import annotations
 
-import numpy as np
+from typing import TYPE_CHECKING
 
 from onnx.reference.op_run import OpRun
 
+if TYPE_CHECKING:
+    import numpy as np
+
 
 class If(OpRun):
     def __init__(self, onnx_node, run_params):  # type: ignore
@@ -31,8 +34,8 @@ class If(OpRun):
         self,
         cond: np.ndarray | np.bool_,
         context=None,
-        else_branch=None,
-        then_branch=None,
+        else_branch=None,  # noqa: ARG002
+        then_branch=None,  # noqa: ARG002
         attributes=None,
     ):
         if cond.size != 1:

onnx/reference/ops/op_instance_normalization.py

@@ -1,7 +1,7 @@
 # Copyright (c) ONNX Project Contributors
 
 # SPDX-License-Identifier: Apache-2.0
-
+from __future__ import annotations
 
 import numpy as np
 

onnx/reference/ops/op_isinf.py

@@ -1,7 +1,7 @@
 # Copyright (c) ONNX Project Contributors
 
 # SPDX-License-Identifier: Apache-2.0
-
+from __future__ import annotations
 
 import numpy as np
 

onnx/reference/ops/op_isnan.py

@@ -1,7 +1,7 @@
 # Copyright (c) ONNX Project Contributors
 
 # SPDX-License-Identifier: Apache-2.0
-
+from __future__ import annotations
 
 import numpy as np
 

onnx/reference/ops/op_layer_normalization.py

@@ -1,9 +1,7 @@
 # Copyright (c) ONNX Project Contributors
 
 # SPDX-License-Identifier: Apache-2.0
-
-
-from typing import Tuple
+from __future__ import annotations
 
 import numpy as np
 
@@ -16,7 +14,7 @@ def _layer_normalization(
     B: np.ndarray,
     axis: int = -1,
     epsilon: float = 1e-5,
-) -> Tuple[np.ndarray, np.ndarray, np.ndarray]:
+) -> tuple[np.ndarray, np.ndarray, np.ndarray]:
     X_shape = X.shape
     X_rank = len(X_shape)
     if axis < 0:

onnx/reference/ops/op_leaky_relu.py

@@ -1,12 +1,15 @@
 # Copyright (c) ONNX Project Contributors
 
 # SPDX-License-Identifier: Apache-2.0
+from __future__ import annotations
 
-
-import numpy as np
+from typing import TYPE_CHECKING
 
 from onnx.reference.ops._op import OpRunUnaryNum
 
+if TYPE_CHECKING:
+    import numpy as np
+
 
 def _leaky_relu(x: np.ndarray, alpha: float) -> np.ndarray:
     sign = (x > 0).astype(x.dtype)

onnx/reference/ops/op_less.py

@@ -1,7 +1,7 @@
 # Copyright (c) ONNX Project Contributors
 
 # SPDX-License-Identifier: Apache-2.0
-
+from __future__ import annotations
 
 import numpy as np
 

onnx/reference/ops/op_less_or_equal.py

@@ -1,7 +1,7 @@
 # Copyright (c) ONNX Project Contributors
 
 # SPDX-License-Identifier: Apache-2.0
-
+from __future__ import annotations
 
 import numpy as np
 

onnx/reference/ops/op_log.py

@@ -1,7 +1,7 @@
 # Copyright (c) ONNX Project Contributors
 
 # SPDX-License-Identifier: Apache-2.0
-
+from __future__ import annotations
 
 import numpy as np
 

onnx/reference/ops/op_log_softmax.py

@@ -1,7 +1,7 @@
 # Copyright (c) ONNX Project Contributors
 
 # SPDX-License-Identifier: Apache-2.0
-
+from __future__ import annotations
 
 import numpy as np
 

onnx/reference/ops/op_loop.py

@@ -1,7 +1,7 @@
 # Copyright (c) ONNX Project Contributors
 
 # SPDX-License-Identifier: Apache-2.0
-
+from __future__ import annotations
 
 import numpy as np
 
@@ -45,7 +45,7 @@ class Loop(OpRun):
             begin = len(loop_inputs) - len(args)
             all_inputs = loop_inputs[begin:]
             for name, val in zip(all_inputs, args):
-                inputs[name] = val
+                inputs[name] = val  # noqa: PERF403
         if context is not None:
             for a in context:
                 inputs[a] = context[a]
@@ -55,7 +55,9 @@ class Loop(OpRun):
         while cond and (M is None or it < M):
             self._log("  -- loop> {%r}", context)
             if len(body.input_names) > 0 and body.input_names[0] is not None:
-                inputs[body.input_names[0]] = np.array(it, dtype=None if M is None else M.dtype)  # type: ignore
+                inputs[body.input_names[0]] = np.array(
+                    it, dtype=None if M is None else M.dtype
+                )  # type: ignore
             if len(body.input_names) > 1 and body.input_names[1] is not None:
                 inputs[body.input_names[1]] = cond
             outputs = self._run_body(inputs, attributes=attributes)  # type: ignore

onnx/reference/ops/op_lp_normalization.py

@@ -1,7 +1,7 @@
 # Copyright (c) ONNX Project Contributors
 
 # SPDX-License-Identifier: Apache-2.0
-
+from __future__ import annotations
 
 import numpy as np
 

onnx/reference/ops/op_lp_pool.py

@@ -1,7 +1,7 @@
 # Copyright (c) ONNX Project Contributors
 
 # SPDX-License-Identifier: Apache-2.0
-
+from __future__ import annotations
 
 import numpy as np
 
@@ -38,4 +38,6 @@ class LpPool(CommonPool):
         )
 
         kernel_element_count = np.prod(kernel_shape)
-        return (np.power(kernel_element_count * power_average[0], 1.0 / p),)
+        return (
+            np.power(kernel_element_count * power_average[0], 1.0 / p).astype(x.dtype),
+        )

onnx/reference/ops/op_lrn.py

@@ -1,7 +1,7 @@
 # Copyright (c) ONNX Project Contributors
 
 # SPDX-License-Identifier: Apache-2.0
-
+from __future__ import annotations
 
 import math
 

onnx/reference/ops/op_lstm.py

@@ -1,9 +1,7 @@
 # Copyright (c) ONNX Project Contributors
 
 # SPDX-License-Identifier: Apache-2.0
-
-
-from typing import Tuple
+from __future__ import annotations
 
 import numpy as np
 
@@ -35,7 +33,7 @@ class CommonLSTM(OpRun):
         C_0: np.ndarray,
         P: np.ndarray,
         num_directions: int,
-    ) -> Tuple[np.ndarray, np.ndarray]:
+    ) -> tuple[np.ndarray, np.ndarray]:
         seq_length = X.shape[0]
         hidden_size = H_0.shape[-1]
         batch_size = X.shape[1]
@@ -85,14 +83,14 @@ class CommonLSTM(OpRun):
         initial_h=None,
         initial_c=None,
         P=None,
-        activation_alpha=None,
-        activation_beta=None,
-        activations=None,
-        clip=None,
-        direction=None,
+        activation_alpha=None,  # noqa: ARG002
+        activation_beta=None,  # noqa: ARG002
+        activations=None,  # noqa: ARG002
+        clip=None,  # noqa: ARG002
+        direction=None,  # noqa: ARG002
         hidden_size=None,
-        input_forget=None,
-        layout=None,
+        input_forget=None,  # noqa: ARG002
+        layout=None,  # noqa: ARG002
     ):
         # TODO: support overridden attributes.
         n_gates = 4

onnx/reference/ops/op_matmul.py

@@ -1,7 +1,7 @@
 # Copyright (c) ONNX Project Contributors
 
 # SPDX-License-Identifier: Apache-2.0
-
+from __future__ import annotations
 
 import numpy as np
 

onnx/reference/ops/op_matmul_integer.py

@@ -1,7 +1,7 @@
 # Copyright (c) ONNX Project Contributors
 
 # SPDX-License-Identifier: Apache-2.0
-
+from __future__ import annotations
 
 import numpy as np
 

onnx/reference/ops/op_max.py

@@ -1,7 +1,7 @@
 # Copyright (c) ONNX Project Contributors
 
 # SPDX-License-Identifier: Apache-2.0
-
+from __future__ import annotations
 
 import numpy as np
 

onnx/reference/ops/op_max_pool.py

@@ -1,7 +1,7 @@
 # Copyright (c) ONNX Project Contributors
 
 # SPDX-License-Identifier: Apache-2.0
-
+from __future__ import annotations
 
 import numpy as np
 
@@ -182,12 +182,12 @@ class MaxPool(CommonPool):
     def _max_pool_1d(  # type: ignore
         self,
         x,
-        auto_pad,
-        ceil_mode,
+        auto_pad,  # noqa: ARG002
+        ceil_mode,  # noqa: ARG002
         dilations,
         kernel_shape,
         new_pads,
-        storage_order,
+        storage_order,  # noqa: ARG002
         strides,
         output_spatial_shape,
     ):
@@ -238,8 +238,8 @@ class MaxPool(CommonPool):
     def _max_pool_2d(  # type: ignore
         self,
         x,
-        auto_pad,
-        ceil_mode,
+        auto_pad,  # noqa: ARG002
+        ceil_mode,  # noqa: ARG002
         dilations,
         kernel_shape,
         new_pads,
@@ -315,8 +315,8 @@ class MaxPool(CommonPool):
     def _max_pool_3d(  # type: ignore
         self,
         x,
-        auto_pad,
-        ceil_mode,
+        auto_pad,  # noqa: ARG002
+        ceil_mode,  # noqa: ARG002
         dilations,
         kernel_shape,
         new_pads,

onnx/reference/ops/op_max_unpool.py

@@ -1,7 +1,7 @@
 # Copyright (c) ONNX Project Contributors
 
 # SPDX-License-Identifier: Apache-2.0
-
+from __future__ import annotations
 
 import numpy as np
 
@@ -9,7 +9,9 @@ from onnx.reference.op_run import OpRun
 
 
 class MaxUnpool(OpRun):
-    def _run(self, X, indices, output_shape=None, kernel_shape=None, pads=None, strides=None):  # type: ignore
+    def _run(
+        self, X, indices, output_shape=None, kernel_shape=None, pads=None, strides=None
+    ):  # type: ignore
         pooling_dims = len(X.shape) - 2
         if pooling_dims > 3:
             raise NotImplementedError(
@@ -28,7 +30,7 @@ class MaxUnpool(OpRun):
         inferred_shape[0] = X.shape[0]
         inferred_shape[1] = X.shape[1]
 
-        for dim in range(0, len(kernel_shape)):
+        for dim in range(len(kernel_shape)):
             inferred_shape[dim + 2] = (
                 (X.shape[dim + 2] - 1) * strides[dim]
                 - (pads[dim] + pads[len(kernel_shape) + dim])

onnx/reference/ops/op_mean.py

@@ -1,7 +1,7 @@
 # Copyright (c) ONNX Project Contributors
 
 # SPDX-License-Identifier: Apache-2.0
-
+from __future__ import annotations
 
 from onnx.reference.op_run import OpRun
 

onnx/reference/ops/op_mel_weight_matrix.py

@@ -1,7 +1,7 @@
 # Copyright (c) ONNX Project Contributors
 
 # SPDX-License-Identifier: Apache-2.0
-
+from __future__ import annotations
 
 import numpy as np
 

onnx/reference/ops/op_min.py

@@ -1,7 +1,7 @@
 # Copyright (c) ONNX Project Contributors
 
 # SPDX-License-Identifier: Apache-2.0
-
+from __future__ import annotations
 
 import numpy as np
 

onnx/reference/ops/op_mod.py

@@ -1,7 +1,7 @@
 # Copyright (c) ONNX Project Contributors
 
 # SPDX-License-Identifier: Apache-2.0
-
+from __future__ import annotations
 
 import numpy as np
 

onnx/reference/ops/op_mul.py

@@ -1,7 +1,7 @@
 # Copyright (c) ONNX Project Contributors
 
 # SPDX-License-Identifier: Apache-2.0
-
+from __future__ import annotations
 
 import numpy as np
 

onnx/reference/ops/op_neg.py

@@ -1,7 +1,7 @@
 # Copyright (c) ONNX Project Contributors
 
 # SPDX-License-Identifier: Apache-2.0
-
+from __future__ import annotations
 
 import numpy as np
 

onnx/reference/ops/op_negative_log_likelihood_loss.py

@@ -1,14 +1,16 @@
 # Copyright (c) ONNX Project Contributors
 
 # SPDX-License-Identifier: Apache-2.0
-
+from __future__ import annotations
 
 import numpy as np
 
 from onnx.reference.op_run import OpRun
 
 
-def _compute_negative_log_likelihood_loss(x, target, weight=None, reduction="mean", ignore_index=None):  # type: ignore
+def _compute_negative_log_likelihood_loss(
+    x, target, weight=None, reduction="mean", ignore_index=None
+):  # type: ignore
     input_shape = x.shape
     if len(input_shape) == 1:
         raise RuntimeError(f"Unsupported shape {input_shape!r}.")

onnx/reference/ops/op_non_max_suppression.py

@@ -1,10 +1,9 @@
 # Copyright (c) ONNX Project Contributors
 
 # SPDX-License-Identifier: Apache-2.0
-
+from __future__ import annotations
 
 import dataclasses
-from typing import Optional, Tuple
 
 import numpy as np
 
@@ -13,20 +12,20 @@ from onnx.reference.op_run import OpRun
 
 @dataclasses.dataclass
 class PrepareContext:
-    boxes_data_: Optional[np.ndarray] = None
+    boxes_data_: np.ndarray | None = None
     boxes_size_: int = 0
-    scores_data_: Optional[np.ndarray] = None
+    scores_data_: np.ndarray | None = None
     scores_size_: int = 0
-    max_output_boxes_per_class_: Optional[np.ndarray] = None
-    score_threshold_: Optional[np.ndarray] = None
-    iou_threshold_: Optional[np.ndarray] = None
+    max_output_boxes_per_class_: np.ndarray | None = None
+    score_threshold_: np.ndarray | None = None
+    iou_threshold_: np.ndarray | None = None
     num_batches_: int = 0
     num_classes_: int = 0
     num_boxes_: int = 0
 
 
 class SelectedIndex:
-    __slots__ = ("batch_index_", "class_index_", "box_index_")
+    __slots__ = ("batch_index_", "box_index_", "class_index_")
 
     def __init__(
         self, batch_index: int = 0, class_index: int = 0, box_index: int = 0
@@ -36,7 +35,7 @@ class SelectedIndex:
         self.box_index_ = box_index
 
 
-def max_min(lhs: float, rhs: float) -> Tuple[float, float]:
+def max_min(lhs: float, rhs: float) -> tuple[float, float]:
     if lhs >= rhs:
         return rhs, lhs
     return lhs, rhs
@@ -133,7 +132,7 @@ class NonMaxSuppression(OpRun):
         max_output_boxes_per_class: int,
         iou_threshold: float,
         score_threshold: float,
-    ) -> Tuple[int, float, float]:
+    ) -> tuple[int, float, float]:
         if pc.max_output_boxes_per_class_ is not None:
             max_output_boxes_per_class = max(pc.max_output_boxes_per_class_[0], 0)
 
@@ -161,7 +160,7 @@ class NonMaxSuppression(OpRun):
             pc.max_output_boxes_per_class_ = max_output_boxes_per_class_tensor
         if iou_threshold_tensor.size != 0:
             pc.iou_threshold_ = iou_threshold_tensor
-        if score_threshold_tensor.size != 0:
+        if score_threshold_tensor is not None and score_threshold_tensor.size != 0:
             pc.score_threshold_ = score_threshold_tensor
 
         pc.boxes_size_ = boxes_tensor.size
@@ -178,13 +177,11 @@ class NonMaxSuppression(OpRun):
         self,
         boxes,
         scores,
-        max_output_boxes_per_class,
-        iou_threshold,
-        score_threshold,
-        center_point_box,
+        max_output_boxes_per_class=None,
+        iou_threshold=None,
+        score_threshold=None,
+        center_point_box=None,
     ):
-        center_point_box = center_point_box or self.center_point_box  # type: ignore
-
         pc = PrepareContext()
         self.prepare_compute(
             pc,
@@ -225,12 +222,12 @@ class NonMaxSuppression(OpRun):
                 if pc.score_threshold_ is not None:
                     for box_index in range(pc.num_boxes_):
                         if class_scores[box_index] > score_threshold:
-                            candidate_boxes.append(
+                            candidate_boxes.append(  # noqa: PERF401
                                 BoxInfo(class_scores[box_index], box_index)
                             )
                 else:
                     for box_index in range(pc.num_boxes_):
-                        candidate_boxes.append(
+                        candidate_boxes.append(  # noqa: PERF401
                             BoxInfo(class_scores[box_index], box_index)
                         )
 

onnx/reference/ops/op_non_zero.py

@@ -1,7 +1,7 @@
 # Copyright (c) ONNX Project Contributors
 
 # SPDX-License-Identifier: Apache-2.0
-
+from __future__ import annotations
 
 import numpy as np