[pull] main from kornia:main

.pre-commit-config.yaml

@@ -15,7 +15,7 @@ repos:
         exclude: ^$|.devcontainer
 
   - repo: https://github.com/astral-sh/ruff-pre-commit
-    rev: v0.8.2
+    rev: v0.8.3
     hooks:
       - id: ruff
         args: [--fix, --exit-non-zero-on-fix]

kornia/augmentation/auto/base.py

@@ -56,6 +56,7 @@ def get_transformation_matrix(
             params: params for the sequence.
             recompute: if to recompute the transformation matrix according to the params.
                 default: False.
+            extra_args: Optional dictionary of extra arguments with specific options for different input types.
 
         """
         if params is None:

kornia/augmentation/auto/operations/policy.py

@@ -58,6 +58,7 @@ def get_transformation_matrix(
             params: params for the sequence.
             recompute: if to recompute the transformation matrix according to the params.
                 default: False.
+            extra_args: Optional dictionary of extra arguments with specific options for different input types.
 
         """
         if params is None:

kornia/augmentation/container/augment.py

@@ -480,8 +480,10 @@ def __call__(
         """Overwrite the __call__ function to handle various inputs.
 
         Args:
+            inputs: Inputs to operate on.
             input_names_to_handle: List of input names to convert, if None, handle all inputs.
             output_type: Desired output type ('tensor', 'numpy', or 'pil').
+            kwargs: Additional arguments.
 
         Returns:
             Callable: Decorated function with converted input and output types.

kornia/augmentation/container/base.py

@@ -185,7 +185,7 @@ def get_transformation_matrix(
             params: params for the sequence.
             recompute: if to recompute the transformation matrix according to the params.
                 default: False.
-
+            extra_args: Optional dictionary of extra arguments with specific options for different input types.
         """
         raise NotImplementedError
 

kornia/augmentation/container/image.py

@@ -248,7 +248,7 @@ def get_transformation_matrix(
             params: params for the sequence.
             recompute: if to recompute the transformation matrix according to the params.
                 default: False.
-
+            extra_args: Optional dictionary of extra arguments with specific options for different input types.
         """
         if params is None:
             raise NotImplementedError("requires params to be provided.")
@@ -326,8 +326,10 @@ def __call__(
         """Overwrite the __call__ function to handle various inputs.
 
         Args:
+            inputs: Inputs to operate on.
             input_names_to_handle: List of input names to convert, if None, handle all inputs.
             output_type: Desired output type ('tensor', 'numpy', or 'pil').
+            kwargs: Additional arguments.
 
         Returns:
             Callable: Decorated function with converted input and output types.

kornia/augmentation/container/ops.py

@@ -50,7 +50,7 @@ def transform(cls, input: T, module: Module, param: ParamItem, extra_args: Optio
             module: any torch Module but only kornia augmentation modules will count
                 to apply transformations.
             param: the corresponding parameters to the module.
-
+            extra_args: Optional dictionary of extra arguments with specific options for different input types.
         """
         raise NotImplementedError
 
@@ -64,7 +64,7 @@ def inverse(cls, input: T, module: Module, param: ParamItem, extra_args: Optiona
             module: any torch Module but only kornia augmentation modules will count
                 to apply transformations.
             param: the corresponding parameters to the module.
-
+            extra_args: Optional dictionary of extra arguments with specific options for different input types.
         """
         raise NotImplementedError
 
@@ -255,7 +255,7 @@ def transform(
             module: any torch Module but only kornia augmentation modules will count
                 to apply transformations.
             param: the corresponding parameters to the module.
-
+            extra_args: Optional dictionary of extra arguments with specific options for different input types.
         """
         if extra_args is None:
             extra_args = {}
@@ -304,7 +304,7 @@ def transform_list(
             module: any torch Module but only kornia augmentation modules will count
                 to apply transformations.
             param: the corresponding parameters to the module.
-
+            extra_args: Optional dictionary of extra arguments with specific options for different input types.
         """
         if extra_args is None:
             extra_args = {}
@@ -369,7 +369,7 @@ def inverse(
             module: any torch Module but only kornia augmentation modules will count
                 to apply transformations.
             param: the corresponding parameters to the module.
-
+            extra_args: Optional dictionary of extra arguments with specific options for different input types.
         """
         if extra_args is None:
             extra_args = {}
@@ -417,7 +417,7 @@ def transform(
             module: any torch Module but only kornia augmentation modules will count
                 to apply transformations.
             param: the corresponding parameters to the module.
-
+            extra_args: Optional dictionary of extra arguments with specific options for different input types.
         """
         if extra_args is None:
             extra_args = {}
@@ -463,7 +463,7 @@ def inverse(
             module: any torch Module but only kornia augmentation modules will count
                 to apply transformations.
             param: the corresponding parameters to the module.
-
+            extra_args: Optional dictionary of extra arguments with specific options for different input types.
         """
         if extra_args is None:
             extra_args = {}
@@ -514,7 +514,7 @@ def transform(
             module: any torch Module but only kornia augmentation modules will count
                 to apply transformations.
             param: the corresponding parameters to the module.
-
+            extra_args: Optional dictionary of extra arguments with specific options for different input types.
         """
         if extra_args is None:
             extra_args = {}
@@ -561,7 +561,7 @@ def inverse(
             module: any torch Module but only kornia augmentation modules will count
                 to apply transformations.
             param: the corresponding parameters to the module.
-
+            extra_args: Optional dictionary of extra arguments with specific options for different input types.
         """
         if extra_args is None:
             extra_args = {}

kornia/augmentation/container/video.py

@@ -253,7 +253,8 @@ def transform_boxes(  # type: ignore[override]
         Args:
             input: tensor with shape :math:`(B, T, N, 4, 2)`.
                 If input is a `Keypoints` type, the internal shape is :math:`(B * T, N, 4, 2)`.
-
+            params: params for the sequence.
+            extra_args: Optional dictionary of extra arguments with specific options for different input types.
         """
         if isinstance(input, Tensor):
             batchsize, frame_num = input.size(0), input.size(1)
@@ -272,7 +273,8 @@ def inverse_boxes(  # type: ignore[override]
         Args:
             input: tensor with shape :math:`(B, T, N, 4, 2)`.
                 If input is a `Keypoints` type, the internal shape is :math:`(B * T, N, 4, 2)`.
-
+            params: params for the sequence.
+            extra_args: Optional dictionary of extra arguments with specific options for different input types.
         """
         if isinstance(input, Tensor):
             batchsize, frame_num = input.size(0), input.size(1)
@@ -291,7 +293,8 @@ def transform_keypoints(  # type: ignore[override]
         Args:
             input: tensor with shape :math:`(B, T, N, 2)`.
                 If input is a `Keypoints` type, the internal shape is :math:`(B * T, N, 2)`.
-
+            params: params for the sequence.
+            extra_args: Optional dictionary of extra arguments with specific options for different input types.
         """
         if isinstance(input, Tensor):
             batchsize, frame_num = input.size(0), input.size(1)
@@ -310,7 +313,8 @@ def inverse_keypoints(  # type: ignore[override]
         Args:
             input: tensor with shape :math:`(B, T, N, 2)`.
                 If input is a `Keypoints` type, the internal shape is :math:`(B * T, N, 2)`.
-
+            params: params for the sequence.
+            extra_args: Optional dictionary of extra arguments with specific options for different input types.
         """
         if isinstance(input, Tensor):
             frame_num, batchsize = input.size(0), input.size(1)

kornia/augmentation/utils/helpers.py

@@ -211,8 +211,9 @@ def _transform_output_shape(
     r"""Collapse the broadcasted batch dimensions an input tensor to be the specified shape.
 
     Args:
-        input: Tensor
+        output: Tensor
         shape: List/tuple of int
+        reference_shape: Tensor representation of shape to control which dimensions are collapsed.
 
     Returns:
         Tensor

kornia/contrib/extract_patches.py

@@ -430,7 +430,7 @@ def extract_tensor_patches(
         window_size: the size of the sliding window and the output patch size.
         stride: stride of the sliding window.
         padding: Zero-padding added to both side of the input.
-        allow_auto_adding: whether to allow automatic padding if the window and stride do not fit into the image.
+        allow_auto_padding: whether to allow automatic padding if the window and stride do not fit into the image.
 
     Returns:
         the tensor with the extracted patches with shape :math:`(B, N, C, H_{out}, W_{out})`.

kornia/contrib/models/common.py

@@ -128,7 +128,7 @@ def window_unpartition(windows: Tensor, window_size: int, pad_hw: tuple[int, int
     """Window unpartition into original sequences and removing padding.
 
     Args:
-        x: input tokens with [B * num_windows, window_size, window_size, C].
+        windows: input tokens with [B * num_windows, window_size, window_size, C].
         window_size: window size.
         pad_hw: padded height and width (Hp, Wp).
         hw: original height and width (H, W) before padding.

kornia/contrib/models/rt_detr/model.py

@@ -81,6 +81,7 @@ def from_name(model_name: str, num_classes: int = 80) -> RTDETRConfig:
 
         Args:
             model_name: 'rtdetr_r18vd', 'rtdetr_r34vd', 'rtdetr_r50vd_m', 'rtdetr_r50vd', 'rtdetr_r101vd'.
+            num_classes: Number of classes to detect.
 
         """
         if model_name == "rtdetr_r18vd":
@@ -107,8 +108,8 @@ def __init__(self, backbone: ResNetD | PPHGNetV2, encoder: HybridEncoder, decode
 
         Args:
             backbone: backbone network for feature extraction.
-            neck: neck network for feature fusion.
-            head: head network to decode features into detection results.
+            encoder: neck network for feature fusion.
+            decoder: head network to decode features into detection results.
 
         """
         super().__init__()
@@ -255,6 +256,7 @@ def from_name(model_name: str, num_classes: int = 80) -> RTDETR:
 
         Args:
             model_name: 'rtdetr_r18vd', 'rtdetr_r34vd', 'rtdetr_r50vd_m', 'rtdetr_r50vd', 'rtdetr_r101vd'.
+            num_classes: number of classes to detect.
 
         """
         model = RTDETR.from_config(RTDETRConfig.from_name(model_name, num_classes))

kornia/contrib/models/sam/architecture/image_encoder.py

@@ -49,6 +49,7 @@ def __init__(
         depth: Depth of ViT.
         num_heads: Number of attention heads in each ViT block.
         mlp_ratio: Ratio of mlp hidden dim to embedding dim.
+        out_chans: Number of output channels.
         qkv_bias: If True, add a learnable bias to query, key, value.
         norm_layer: Normalization layer.
         act_layer: Activation layer.
@@ -194,7 +195,7 @@ def __init__(
         dim: Number of input channels.
         num_heads: Number of attention heads.
         qkv_bias:  If True, add a learnable bias to query, key, value.
-        rel_pos: If True, add relative positional embeddings to the attention map.
+        use_rel_pos: If True, add relative positional embeddings to the attention map.
         rel_pos_zero_init: If True, zero initialize relative positional parameters.
         input_size: Input resolution for calculating the relative positional parameter size.
 

kornia/contrib/models/sam/architecture/transformer.py

@@ -34,6 +34,7 @@ def __init__(
             num_heads: the number of heads for multihead attention. Must divide embedding_dim
             mlp_dim: the channel dimension internal to the MLP block
             activation: the activation to use in the MLP block
+            attention_downsample_rate: downsampling rate from embedding dimension
 
         """  # noqa: D205
         super().__init__()
@@ -117,6 +118,7 @@ def __init__(
             mlp_dim: the hidden dimension of the mlp block
             activation: the activation of the mlp block
             skip_first_layer_pe: skip the PE on the first layer
+            attention_downsample_rate: downsampling rate from embedding dimension
 
         """
         super().__init__()

kornia/contrib/visual_prompter.py

@@ -119,6 +119,8 @@ def set_image(self, image: Tensor, mean: Optional[Tensor] = None, std: Optional[
             image: RGB image. Normally images with range of [0-1], the model preprocess normalize the
                    pixel values with the mean and std defined in its initialization. Expected to be into a float32
                    dtype. Shape :math:`(3, H, W)`.
+            mean: mean value of dataset for normalization.
+            std: standard deviation of dataset for normalization.
 
         """
         KORNIA_CHECK_SHAPE(image, ["3", "H", "W"])
@@ -236,7 +238,7 @@ def predict(
         Args:
             keypoints: Point prompts to the model. Each point is in (X,Y) in pixels. Shape :math:`(K, N, 2)`. Where
                        `N` is the number of points and `K` the number of prompts.
-            keypoint_labels: Labels for the point prompts. 1 indicates a foreground point and 0 indicates a background
+            keypoints_labels: Labels for the point prompts. 1 indicates a foreground point and 0 indicates a background
                              point. Shape :math:`(K, N)`. Where `N` is the number of points, and `K` the number of
                              prompts.
             boxes: A box prompt to the model. If a tensor, should be in a xyxy mode. Shape :math:`(K, 4)`

kornia/core/check.py

@@ -235,7 +235,7 @@ def KORNIA_CHECK_SAME_DEVICES(tensors: list[Tensor], msg: Optional[str] = None,
     """Check whether a list provided tensors live in the same device.
 
     Args:
-        x: a list of tensors.
+        tensors: a list of tensors.
         msg: message to show in the exception.
         raises: bool indicating whether an exception should be raised upon failure.
 

kornia/core/mixin/image_module.py

@@ -190,6 +190,7 @@ def show(self, n_row: Optional[int] = None, backend: str = "pil", display: bool
         Args:
             n_row: Number of images displayed in each row of the grid.
             backend: visualization backend. Only PIL is supported now.
+            display: Whether or not to show the image.
 
         """
         if self._output_image is None: