dgenerate-ultralytics-headless 8.3.194__py3-none-any.whl → 8.3.195__py3-none-any.whl

ultralytics/models/sam/modules/encoders.py

@@ -83,7 +83,7 @@ class ImageEncoderViT(nn.Module):
             use_rel_pos (bool): If True, adds relative positional embeddings to attention maps.
             rel_pos_zero_init (bool): If True, initializes relative positional parameters to zero.
             window_size (int): Size of attention window for windowed attention blocks.
-            global_attn_indexes (Tuple[int, ...]): Indices of blocks that use global attention.
+            global_attn_indexes (tuple[int, ...]): Indices of blocks that use global attention.
 
         Examples:
             >>> encoder = ImageEncoderViT(img_size=224, patch_size=16, embed_dim=768, depth=12, num_heads=12)
@@ -161,13 +161,13 @@ class PromptEncoder(nn.Module):
 
     Attributes:
         embed_dim (int): Dimension of the embeddings.
-        input_image_size (Tuple[int, int]): Size of the input image as (H, W).
-        image_embedding_size (Tuple[int, int]): Spatial size of the image embedding as (H, W).
+        input_image_size (tuple[int, int]): Size of the input image as (H, W).
+        image_embedding_size (tuple[int, int]): Spatial size of the image embedding as (H, W).
         pe_layer (PositionEmbeddingRandom): Module for random position embedding.
         num_point_embeddings (int): Number of point embeddings for different types of points.
         point_embeddings (nn.ModuleList): List of point embeddings.
         not_a_point_embed (nn.Embedding): Embedding for points that are not part of any label.
-        mask_input_size (Tuple[int, int]): Size of the input mask.
+        mask_input_size (tuple[int, int]): Size of the input mask.
         mask_downscaling (nn.Sequential): Neural network for downscaling the mask.
         no_mask_embed (nn.Embedding): Embedding for cases where no mask is provided.
 
@@ -198,8 +198,8 @@ class PromptEncoder(nn.Module):
 
         Args:
             embed_dim (int): The dimension of the embeddings.
-            image_embedding_size (Tuple[int, int]): The spatial size of the image embedding as (H, W).
-            input_image_size (Tuple[int, int]): The padded size of the input image as (H, W).
+            image_embedding_size (tuple[int, int]): The spatial size of the image embedding as (H, W).
+            input_image_size (tuple[int, int]): The padded size of the input image as (H, W).
             mask_in_chans (int): The number of hidden channels used for encoding input masks.
             activation (Type[nn.Module]): The activation function to use when encoding input masks.
 
@@ -310,7 +310,7 @@ class PromptEncoder(nn.Module):
         Embed different types of prompts, returning both sparse and dense embeddings.
 
         Args:
-            points (Tuple[torch.Tensor, torch.Tensor] | None): Point coordinates and labels to embed. The first
+            points (tuple[torch.Tensor, torch.Tensor] | None): Point coordinates and labels to embed. The first
                 tensor contains coordinates with shape (B, N, 2), and the second tensor contains labels with
                 shape (B, N).
             boxes (torch.Tensor | None): Boxes to embed with shape (B, M, 2, 2), where M is the number of boxes.
@@ -522,10 +522,10 @@ class FpnNeck(nn.Module):
     Attributes:
         position_encoding (PositionEmbeddingSine): Sinusoidal positional encoding module.
         convs (nn.ModuleList): List of convolutional layers for each backbone level.
-        backbone_channel_list (List[int]): List of channel dimensions from the backbone.
+        backbone_channel_list (list[int]): List of channel dimensions from the backbone.
         fpn_interp_model (str): Interpolation mode for FPN feature resizing.
         fuse_type (str): Type of feature fusion, either 'sum' or 'avg'.
-        fpn_top_down_levels (List[int]): Levels to have top-down features in outputs.
+        fpn_top_down_levels (list[int]): Levels to have top-down features in outputs.
 
     Methods:
         forward: Perform forward pass through the FPN neck.
@@ -558,13 +558,13 @@ class FpnNeck(nn.Module):
 
         Args:
             d_model (int): Dimension of the model.
-            backbone_channel_list (List[int]): List of channel dimensions from the backbone.
+            backbone_channel_list (list[int]): List of channel dimensions from the backbone.
             kernel_size (int): Kernel size for the convolutional layers.
             stride (int): Stride for the convolutional layers.
             padding (int): Padding for the convolutional layers.
             fpn_interp_model (str): Interpolation mode for FPN feature resizing.
             fuse_type (str): Type of feature fusion, either 'sum' or 'avg'.
-            fpn_top_down_levels (Optional[List[int]]): Levels to have top-down features in outputs.
+            fpn_top_down_levels (Optional[list[int]]): Levels to have top-down features in outputs.
 
         Examples:
             >>> backbone_channels = [64, 128, 256, 512]
@@ -610,12 +610,12 @@ class FpnNeck(nn.Module):
         and top-down feature fusion. It generates output feature maps and corresponding positional encodings.
 
         Args:
-            xs (List[torch.Tensor]): List of input tensors from the backbone, each with shape (B, C, H, W).
+            xs (list[torch.Tensor]): List of input tensors from the backbone, each with shape (B, C, H, W).
 
         Returns:
-            out (List[torch.Tensor]): List of output feature maps after FPN processing, each with shape
+            out (list[torch.Tensor]): List of output feature maps after FPN processing, each with shape
                 (B, d_model, H, W).
-            pos (List[torch.Tensor]): List of positional encodings corresponding to each output feature map.
+            pos (list[torch.Tensor]): List of positional encodings corresponding to each output feature map.
 
         Examples:
             >>> fpn_neck = FpnNeck(d_model=256, backbone_channel_list=[64, 128, 256, 512])
@@ -664,18 +664,18 @@ class Hiera(nn.Module):
     with optional pooling and global attention mechanisms.
 
     Attributes:
-        window_spec (Tuple[int, ...]): Window sizes for each stage.
-        q_stride (Tuple[int, int]): Downsampling stride between stages.
-        stage_ends (List[int]): Indices of the last block in each stage.
-        q_pool_blocks (List[int]): Indices of blocks where pooling is applied.
+        window_spec (tuple[int, ...]): Window sizes for each stage.
+        q_stride (tuple[int, int]): Downsampling stride between stages.
+        stage_ends (list[int]): Indices of the last block in each stage.
+        q_pool_blocks (list[int]): Indices of blocks where pooling is applied.
         return_interm_layers (bool): Whether to return intermediate layer outputs.
         patch_embed (PatchEmbed): Module for patch embedding.
-        global_att_blocks (Tuple[int, ...]): Indices of blocks with global attention.
-        window_pos_embed_bkg_spatial_size (Tuple[int, int]): Spatial size for window positional embedding background.
+        global_att_blocks (tuple[int, ...]): Indices of blocks with global attention.
+        window_pos_embed_bkg_spatial_size (tuple[int, int]): Spatial size for window positional embedding background.
         pos_embed (nn.Parameter): Positional embedding for the background.
         pos_embed_window (nn.Parameter): Positional embedding for the window.
         blocks (nn.ModuleList): List of MultiScaleBlock modules.
-        channel_list (List[int]): List of output channel dimensions for each stage.
+        channel_list (list[int]): List of output channel dimensions for each stage.
 
     Methods:
         _get_pos_embed: Generate positional embeddings by interpolating and combining window and background embeddings.
@@ -727,13 +727,13 @@ class Hiera(nn.Module):
             num_heads (int): Initial number of attention heads.
             drop_path_rate (float): Stochastic depth rate.
             q_pool (int): Number of query pooling stages.
-            q_stride (Tuple[int, int]): Downsampling stride between stages.
-            stages (Tuple[int, ...]): Number of blocks per stage.
+            q_stride (tuple[int, int]): Downsampling stride between stages.
+            stages (tuple[int, ...]): Number of blocks per stage.
             dim_mul (float): Dimension multiplier factor at stage transitions.
             head_mul (float): Head multiplier factor at stage transitions.
-            window_pos_embed_bkg_spatial_size (Tuple[int, int]): Spatial size for window positional embedding background.
-            window_spec (Tuple[int, ...]): Window sizes for each stage when not using global attention.
-            global_att_blocks (Tuple[int, ...]): Indices of blocks that use global attention.
+            window_pos_embed_bkg_spatial_size (tuple[int, int]): Spatial size for window positional embedding background.
+            window_spec (tuple[int, ...]): Window sizes for each stage when not using global attention.
+            global_att_blocks (tuple[int, ...]): Indices of blocks that use global attention.
             return_interm_layers (bool): Whether to return intermediate layer outputs.
 
         Examples:
@@ -823,7 +823,7 @@ class Hiera(nn.Module):
             x (torch.Tensor): Input tensor with shape (B, C, H, W) representing a batch of images.
 
         Returns:
-            (List[torch.Tensor]): List of feature maps at different scales, each with shape (B, C_i, H_i, W_i), where
+            (list[torch.Tensor]): List of feature maps at different scales, each with shape (B, C_i, H_i, W_i), where
                 C_i is the channel dimension and H_i, W_i are the spatial dimensions at scale i. The list is ordered
                 from highest resolution (fine features) to lowest resolution (coarse features) if return_interm_layers
                 is True, otherwise contains only the final output.

ultralytics/models/sam/modules/sam.py

@@ -68,8 +68,8 @@ class SAMModel(nn.Module):
             image_encoder (ImageEncoderViT): The backbone used to encode the image into image embeddings.
             prompt_encoder (PromptEncoder): Encodes various types of input prompts.
             mask_decoder (MaskDecoder): Predicts masks from the image embeddings and encoded prompts.
-            pixel_mean (List[float]): Mean values for normalizing pixels in the input image.
-            pixel_std (List[float]): Standard deviation values for normalizing pixels in the input image.
+            pixel_mean (list[float]): Mean values for normalizing pixels in the input image.
+            pixel_std (list[float]): Standard deviation values for normalizing pixels in the input image.
 
         Examples:
             >>> image_encoder = ImageEncoderViT(...)
@@ -435,14 +435,14 @@ class SAM2Model(torch.nn.Module):
 
         Args:
             backbone_features (torch.Tensor): Image features with shape (B, C, H, W).
-            point_inputs (Dict[str, torch.Tensor] | None): Dictionary containing point prompts.
+            point_inputs (dict[str, torch.Tensor] | None): Dictionary containing point prompts.
                 'point_coords': Tensor of shape (B, P, 2) with float32 dtype, containing absolute
                     pixel-unit coordinates in (x, y) format for P input points.
                 'point_labels': Tensor of shape (B, P) with int32 dtype, where 1 means positive clicks,
                     0 means negative clicks, and -1 means padding.
             mask_inputs (torch.Tensor | None): Mask of shape (B, 1, H*16, W*16), float or bool, with the
                 same spatial size as the image.
-            high_res_features (List[torch.Tensor] | None): List of two feature maps with shapes
+            high_res_features (list[torch.Tensor] | None): List of two feature maps with shapes
                 (B, C, 4*H, 4*W) and (B, C, 2*H, 2*W) respectively, used as high-resolution feature maps
                 for SAM decoder.
             multimask_output (bool): If True, output 3 candidate masks and their IoU estimates; if False,

ultralytics/models/sam/modules/tiny_encoder.py

@@ -81,7 +81,7 @@ class PatchEmbed(nn.Module):
     effectively downsampling the spatial dimensions while increasing the channel dimension.
 
     Attributes:
-        patches_resolution (Tuple[int, int]): Resolution of the patches after embedding.
+        patches_resolution (tuple[int, int]): Resolution of the patches after embedding.
         num_patches (int): Total number of patches.
         in_chans (int): Number of input channels.
         embed_dim (int): Dimension of the embedding.
@@ -203,7 +203,7 @@ class PatchMerging(nn.Module):
     resolution while potentially increasing channel dimensions.
 
     Attributes:
-        input_resolution (Tuple[int, int]): The input resolution (height, width) of the feature map.
+        input_resolution (tuple[int, int]): The input resolution (height, width) of the feature map.
         dim (int): The input dimension of the feature map.
         out_dim (int): The output dimension after merging and projection.
         act (nn.Module): The activation function used between convolutions.
@@ -225,7 +225,7 @@ class PatchMerging(nn.Module):
         Initialize the PatchMerging module for merging and projecting neighboring patches in feature maps.
 
         Args:
-            input_resolution (Tuple[int, int]): The input resolution (height, width) of the feature map.
+            input_resolution (tuple[int, int]): The input resolution (height, width) of the feature map.
             dim (int): The input dimension of the feature map.
             out_dim (int): The output dimension after merging and projection.
             activation (nn.Module): The activation function used between convolutions.
@@ -267,7 +267,7 @@ class ConvLayer(nn.Module):
 
     Attributes:
         dim (int): Dimensionality of the input and output.
-        input_resolution (Tuple[int, int]): Resolution of the input image.
+        input_resolution (tuple[int, int]): Resolution of the input image.
         depth (int): Number of MBConv layers in the block.
         use_checkpoint (bool): Whether to use gradient checkpointing to save memory.
         blocks (nn.ModuleList): List of MBConv layers.
@@ -301,10 +301,10 @@ class ConvLayer(nn.Module):
 
         Args:
             dim (int): The dimensionality of the input and output.
-            input_resolution (Tuple[int, int]): The resolution of the input image.
+            input_resolution (tuple[int, int]): The resolution of the input image.
             depth (int): The number of MBConv layers in the block.
             activation (nn.Module): Activation function applied after each convolution.
-            drop_path (float | List[float], optional): Drop path rate. Single float or a list of floats for each MBConv.
+            drop_path (float | list[float], optional): Drop path rate. Single float or a list of floats for each MBConv.
             downsample (Optional[nn.Module], optional): Function for downsampling the output. None to skip downsampling.
             use_checkpoint (bool, optional): Whether to use gradient checkpointing to save memory.
             out_dim (Optional[int], optional): The dimensionality of the output. None means it will be the same as `dim`.
@@ -456,7 +456,7 @@ class Attention(torch.nn.Module):
             key_dim (int): The dimensionality of the keys and queries.
             num_heads (int, optional): Number of attention heads.
             attn_ratio (float, optional): Attention ratio, affecting the dimensions of the value vectors.
-            resolution (Tuple[int, int], optional): Spatial resolution of the input feature map.
+            resolution (tuple[int, int], optional): Spatial resolution of the input feature map.
         """
         super().__init__()
 
@@ -530,7 +530,7 @@ class TinyViTBlock(nn.Module):
 
     Attributes:
         dim (int): The dimensionality of the input and output.
-        input_resolution (Tuple[int, int]): Spatial resolution of the input feature map.
+        input_resolution (tuple[int, int]): Spatial resolution of the input feature map.
         num_heads (int): Number of attention heads.
         window_size (int): Size of the attention window.
         mlp_ratio (float): Ratio of MLP hidden dimension to embedding dimension.
@@ -567,7 +567,7 @@ class TinyViTBlock(nn.Module):
 
         Args:
             dim (int): Dimensionality of the input and output features.
-            input_resolution (Tuple[int, int]): Spatial resolution of the input feature map (height, width).
+            input_resolution (tuple[int, int]): Spatial resolution of the input feature map (height, width).
             num_heads (int): Number of attention heads.
             window_size (int, optional): Size of the attention window. Must be greater than 0.
             mlp_ratio (float, optional): Ratio of MLP hidden dimension to embedding dimension.
@@ -674,7 +674,7 @@ class BasicLayer(nn.Module):
 
     Attributes:
         dim (int): The dimensionality of the input and output features.
-        input_resolution (Tuple[int, int]): Spatial resolution of the input feature map.
+        input_resolution (tuple[int, int]): Spatial resolution of the input feature map.
         depth (int): Number of TinyViT blocks in this layer.
         use_checkpoint (bool): Whether to use gradient checkpointing to save memory.
         blocks (nn.ModuleList): List of TinyViT blocks that make up this layer.
@@ -712,13 +712,13 @@ class BasicLayer(nn.Module):
 
         Args:
             dim (int): Dimensionality of the input and output features.
-            input_resolution (Tuple[int, int]): Spatial resolution of the input feature map (height, width).
+            input_resolution (tuple[int, int]): Spatial resolution of the input feature map (height, width).
             depth (int): Number of TinyViT blocks in this layer.
             num_heads (int): Number of attention heads in each TinyViT block.
             window_size (int): Size of the local window for attention computation.
             mlp_ratio (float, optional): Ratio of MLP hidden dimension to embedding dimension.
             drop (float, optional): Dropout rate.
-            drop_path (float | List[float], optional): Stochastic depth rate. Can be a float or a list of floats for each block.
+            drop_path (float | list[float], optional): Stochastic depth rate. Can be a float or a list of floats for each block.
             downsample (nn.Module | None, optional): Downsampling layer at the end of the layer. None to skip downsampling.
             use_checkpoint (bool, optional): Whether to use gradient checkpointing to save memory.
             local_conv_size (int, optional): Kernel size for the local convolution in each TinyViT block.
@@ -778,11 +778,11 @@ class TinyViT(nn.Module):
     Attributes:
         img_size (int): Input image size.
         num_classes (int): Number of classification classes.
-        depths (Tuple[int, int, int, int]): Number of blocks in each stage.
+        depths (tuple[int, int, int, int]): Number of blocks in each stage.
         num_layers (int): Total number of layers in the network.
         mlp_ratio (float): Ratio of MLP hidden dimension to embedding dimension.
         patch_embed (PatchEmbed): Module for patch embedding.
-        patches_resolution (Tuple[int, int]): Resolution of embedded patches.
+        patches_resolution (tuple[int, int]): Resolution of embedded patches.
         layers (nn.ModuleList): List of network layers.
         norm_head (nn.LayerNorm): Layer normalization for the classifier head.
         head (nn.Linear): Linear layer for final classification.
@@ -823,10 +823,10 @@ class TinyViT(nn.Module):
             img_size (int, optional): Size of the input image.
             in_chans (int, optional): Number of input channels.
             num_classes (int, optional): Number of classes for classification.
-            embed_dims (Tuple[int, int, int, int], optional): Embedding dimensions for each stage.
-            depths (Tuple[int, int, int, int], optional): Number of blocks in each stage.
-            num_heads (Tuple[int, int, int, int], optional): Number of attention heads in each stage.
-            window_sizes (Tuple[int, int, int, int], optional): Window sizes for each stage.
+            embed_dims (tuple[int, int, int, int], optional): Embedding dimensions for each stage.
+            depths (tuple[int, int, int, int], optional): Number of blocks in each stage.
+            num_heads (tuple[int, int, int, int], optional): Number of attention heads in each stage.
+            window_sizes (tuple[int, int, int, int], optional): Window sizes for each stage.
             mlp_ratio (float, optional): Ratio of MLP hidden dim to embedding dim.
             drop_rate (float, optional): Dropout rate.
             drop_path_rate (float, optional): Stochastic depth rate.

ultralytics/models/sam/modules/utils.py

@@ -14,12 +14,12 @@ def select_closest_cond_frames(frame_idx: int, cond_frame_outputs: dict[int, Any
 
     Args:
         frame_idx (int): Current frame index.
-        cond_frame_outputs (Dict[int, Any]): Dictionary of conditioning frame outputs keyed by frame indices.
+        cond_frame_outputs (dict[int, Any]): Dictionary of conditioning frame outputs keyed by frame indices.
         max_cond_frame_num (int): Maximum number of conditioning frames to select.
 
     Returns:
-        selected_outputs (Dict[int, Any]): Selected items from cond_frame_outputs.
-        unselected_outputs (Dict[int, Any]): Items not selected from cond_frame_outputs.
+        selected_outputs (dict[int, Any]): Selected items from cond_frame_outputs.
+        unselected_outputs (dict[int, Any]): Items not selected from cond_frame_outputs.
 
     Examples:
         >>> frame_idx = 5
@@ -229,7 +229,7 @@ def window_partition(x: torch.Tensor, window_size: int):
 
     Returns:
         windows (torch.Tensor): Partitioned windows with shape (B * num_windows, window_size, window_size, C).
-        padded_h_w (Tuple[int, int]): Padded height and width before partition.
+        padded_h_w (tuple[int, int]): Padded height and width before partition.
 
     Examples:
         >>> x = torch.randn(1, 16, 16, 3)
@@ -262,8 +262,8 @@ def window_unpartition(windows: torch.Tensor, window_size: int, pad_hw: tuple[in
             window_size, C), where B is the batch size, num_windows is the number of windows, window_size is
             the size of each window, and C is the number of channels.
         window_size (int): Size of each window.
-        pad_hw (Tuple[int, int]): Padded height and width (Hp, Wp) of the input before windowing.
-        hw (Tuple[int, int]): Original height and width (H, W) of the input before padding and windowing.
+        pad_hw (tuple[int, int]): Padded height and width (Hp, Wp) of the input before windowing.
+        hw (tuple[int, int]): Original height and width (H, W) of the input before padding and windowing.
 
     Returns:
         (torch.Tensor): Unpartitioned sequences with shape (B, H, W, C), where B is the batch size, H and W
@@ -350,8 +350,8 @@ def add_decomposed_rel_pos(
         q (torch.Tensor): Query tensor in the attention layer with shape (B, q_h * q_w, C).
         rel_pos_h (torch.Tensor): Relative position embeddings for height axis with shape (Lh, C).
         rel_pos_w (torch.Tensor): Relative position embeddings for width axis with shape (Lw, C).
-        q_size (Tuple[int, int]): Spatial sequence size of query q as (q_h, q_w).
-        k_size (Tuple[int, int]): Spatial sequence size of key k as (k_h, k_w).
+        q_size (tuple[int, int]): Spatial sequence size of query q as (q_h, q_w).
+        k_size (tuple[int, int]): Spatial sequence size of key k as (k_h, k_w).
 
     Returns:
         (torch.Tensor): Updated attention map with added relative positional embeddings, shape