RetinaNet object detection.

torchvision/models/detection/__init__.py

@@ -1,3 +1,4 @@
 from .faster_rcnn import *
 from .mask_rcnn import *
 from .keypoint_rcnn import *
+from .retinanet import *

torchvision/models/detection/anchor_utils.py

@@ -0,0 +1,159 @@
+# Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved.
+import torch
+from torch import nn
+
+from torch.jit.annotations import List, Optional, Dict
+from .image_list import ImageList
+
+
+class AnchorGenerator(nn.Module):
+    """
+    Module that generates anchors for a set of feature maps and
+    image sizes.
+
+    The module support computing anchors at multiple sizes and aspect ratios
+    per feature map. This module assumes aspect ratio = height / width for
+    each anchor.
+
+    sizes and aspect_ratios should have the same number of elements, and it should
+    correspond to the number of feature maps.
+
+    sizes[i] and aspect_ratios[i] can have an arbitrary number of elements,
+    and AnchorGenerator will output a set of sizes[i] * aspect_ratios[i] anchors
+    per spatial location for feature map i.
+
+    Arguments:
+        sizes (Tuple[Tuple[int]]):
+        aspect_ratios (Tuple[Tuple[float]]):
+    """
+
+    __annotations__ = {
+        "cell_anchors": Optional[List[torch.Tensor]],
+        "_cache": Dict[str, List[torch.Tensor]]
+    }
+
+    def __init__(
+        self,
+        sizes=((128, 256, 512),),
+        aspect_ratios=((0.5, 1.0, 2.0),),
+    ):
+        super(AnchorGenerator, self).__init__()
+
+        if not isinstance(sizes[0], (list, tuple)):
+            # TODO change this
+            sizes = tuple((s,) for s in sizes)
+        if not isinstance(aspect_ratios[0], (list, tuple)):
+            aspect_ratios = (aspect_ratios,) * len(sizes)
+
+        assert len(sizes) == len(aspect_ratios)
+
+        self.sizes = sizes
+        self.aspect_ratios = aspect_ratios
+        self.cell_anchors = None
+        self._cache = {}
+
+    # TODO: https://github.com/pytorch/pytorch/issues/26792
+    # For every (aspect_ratios, scales) combination, output a zero-centered anchor with those values.
+    # (scales, aspect_ratios) are usually an element of zip(self.scales, self.aspect_ratios)
+    # This method assumes aspect ratio = height / width for an anchor.
+    def generate_anchors(self, scales, aspect_ratios, dtype=torch.float32, device="cpu"):
+        # type: (List[int], List[float], int, Device) -> Tensor  # noqa: F821
+        scales = torch.as_tensor(scales, dtype=dtype, device=device)
+        aspect_ratios = torch.as_tensor(aspect_ratios, dtype=dtype, device=device)
+        h_ratios = torch.sqrt(aspect_ratios)
+        w_ratios = 1 / h_ratios
+
+        ws = (w_ratios[:, None] * scales[None, :]).view(-1)
+        hs = (h_ratios[:, None] * scales[None, :]).view(-1)
+
+        base_anchors = torch.stack([-ws, -hs, ws, hs], dim=1) / 2
+        return base_anchors.round()
+
+    def set_cell_anchors(self, dtype, device):
+        # type: (int, Device) -> None  # noqa: F821
+        if self.cell_anchors is not None:
+            cell_anchors = self.cell_anchors
+            assert cell_anchors is not None
+            # suppose that all anchors have the same device
+            # which is a valid assumption in the current state of the codebase
+            if cell_anchors[0].device == device:
+                return
+
+        cell_anchors = [
+            self.generate_anchors(
+                sizes,
+                aspect_ratios,
+                dtype,
+                device
+            )
+            for sizes, aspect_ratios in zip(self.sizes, self.aspect_ratios)
+        ]
+        self.cell_anchors = cell_anchors
+
+    def num_anchors_per_location(self):
+        return [len(s) * len(a) for s, a in zip(self.sizes, self.aspect_ratios)]
+
+    # For every combination of (a, (g, s), i) in (self.cell_anchors, zip(grid_sizes, strides), 0:2),
+    # output g[i] anchors that are s[i] distance apart in direction i, with the same dimensions as a.
+    def grid_anchors(self, grid_sizes, strides):
+        # type: (List[List[int]], List[List[Tensor]]) -> List[Tensor]
+        anchors = []
+        cell_anchors = self.cell_anchors
+        assert cell_anchors is not None
+        assert len(grid_sizes) == len(strides) == len(cell_anchors)
+
+        for size, stride, base_anchors in zip(
+            grid_sizes, strides, cell_anchors
+        ):
+            grid_height, grid_width = size
+            stride_height, stride_width = stride
+            device = base_anchors.device
+
+            # For output anchor, compute [x_center, y_center, x_center, y_center]
+            shifts_x = torch.arange(
+                0, grid_width, dtype=torch.float32, device=device
+            ) * stride_width
+            shifts_y = torch.arange(
+                0, grid_height, dtype=torch.float32, device=device
+            ) * stride_height
+            shift_y, shift_x = torch.meshgrid(shifts_y, shifts_x)
+            shift_x = shift_x.reshape(-1)
+            shift_y = shift_y.reshape(-1)
+            shifts = torch.stack((shift_x, shift_y, shift_x, shift_y), dim=1)
+
+            # For every (base anchor, output anchor) pair,
+            # offset each zero-centered base anchor by the center of the output anchor.
+            anchors.append(
+                (shifts.view(-1, 1, 4) + base_anchors.view(1, -1, 4)).reshape(-1, 4)
+            )
+
+        return anchors
+
+    def cached_grid_anchors(self, grid_sizes, strides):
+        # type: (List[List[int]], List[List[Tensor]]) -> List[Tensor]
+        key = str(grid_sizes) + str(strides)
+        if key in self._cache:
+            return self._cache[key]
+        anchors = self.grid_anchors(grid_sizes, strides)
+        self._cache[key] = anchors
+        return anchors
+
+    def forward(self, image_list, feature_maps):
+        # type: (ImageList, List[Tensor]) -> List[Tensor]
+        grid_sizes = list([feature_map.shape[-2:] for feature_map in feature_maps])
+        image_size = image_list.tensors.shape[-2:]
+        dtype, device = feature_maps[0].dtype, feature_maps[0].device
+        strides = [[torch.tensor(image_size[0] // g[0], dtype=torch.int64, device=device),
+                    torch.tensor(image_size[1] // g[1], dtype=torch.int64, device=device)] for g in grid_sizes]
+        self.set_cell_anchors(dtype, device)
+        anchors_over_all_feature_maps = self.cached_grid_anchors(grid_sizes, strides)
+        anchors = torch.jit.annotate(List[List[torch.Tensor]], [])
+        for i, (image_height, image_width) in enumerate(image_list.image_sizes):
+            anchors_in_image = []
+            for anchors_per_feature_map in anchors_over_all_feature_maps:
+                anchors_in_image.append(anchors_per_feature_map)
+            anchors.append(anchors_in_image)
+        anchors = [torch.cat(anchors_per_image) for anchors_per_image in anchors]
+        # Clear the cache in case that memory leaks.
+        self._cache.clear()
+        return anchors

torchvision/models/detection/backbone_utils.py

@@ -25,13 +25,17 @@ class BackboneWithFPN(nn.Module):
     Attributes:
         out_channels (int): the number of channels in the FPN
     """
-    def __init__(self, backbone, return_layers, in_channels_list, out_channels):
+    def __init__(self, backbone, return_layers, in_channels_list, out_channels, extra_blocks=None):
         super(BackboneWithFPN, self).__init__()
+
+        if extra_blocks is None:
+            extra_blocks = LastLevelMaxPool()
+
         self.body = IntermediateLayerGetter(backbone, return_layers=return_layers)
         self.fpn = FeaturePyramidNetwork(
             in_channels_list=in_channels_list,
             out_channels=out_channels,
-            extra_blocks=LastLevelMaxPool(),
+            extra_blocks=extra_blocks,
         )
         self.out_channels = out_channels
 
@@ -41,7 +45,16 @@ def forward(self, x):
         return x
 
 
-def resnet_fpn_backbone(backbone_name, pretrained, norm_layer=misc_nn_ops.FrozenBatchNorm2d, trainable_layers=3):
+def resnet_fpn_backbone(
+    backbone_name,
+    pretrained,
+    norm_layer=misc_nn_ops.FrozenBatchNorm2d,
+    trainable_layers=3,
+    extra_blocks=None
+):
+    backbone = resnet.__dict__[backbone_name](
+        pretrained=pretrained,
+        norm_layer=norm_layer)
     """
     Constructs a specified ResNet backbone with FPN on top. Freezes the specified number of layers in the backbone.
 
@@ -82,6 +95,9 @@ def resnet_fpn_backbone(backbone_name, pretrained, norm_layer=misc_nn_ops.Frozen
         if all([not name.startswith(layer) for layer in layers_to_train]):
             parameter.requires_grad_(False)
 
+    if extra_blocks is None:
+        extra_blocks = LastLevelMaxPool()
+
     return_layers = {'layer1': '0', 'layer2': '1', 'layer3': '2', 'layer4': '3'}
 
     in_channels_stage2 = backbone.inplanes // 8
@@ -92,4 +108,4 @@ def resnet_fpn_backbone(backbone_name, pretrained, norm_layer=misc_nn_ops.Frozen
         in_channels_stage2 * 8,
     ]
     out_channels = 256
-    return BackboneWithFPN(backbone, return_layers, in_channels_list, out_channels)
+    return BackboneWithFPN(backbone, return_layers, in_channels_list, out_channels, extra_blocks=extra_blocks)

torchvision/models/detection/faster_rcnn.py

@@ -9,8 +9,9 @@
 
 from ..utils import load_state_dict_from_url
 
+from .anchor_utils import AnchorGenerator
 from .generalized_rcnn import GeneralizedRCNN
-from .rpn import AnchorGenerator, RPNHead, RegionProposalNetwork
+from .rpn import RPNHead, RegionProposalNetwork
 from .roi_heads import RoIHeads
 from .transform import GeneralizedRCNNTransform
 from .backbone_utils import resnet_fpn_backbone

torchvision/models/detection/keypoint_rcnn.py

@@ -103,7 +103,7 @@ class KeypointRCNN(FasterRCNN):
         >>> import torch
         >>> import torchvision
         >>> from torchvision.models.detection import KeypointRCNN
-        >>> from torchvision.models.detection.rpn import AnchorGenerator
+        >>> from torchvision.models.detection.anchor_utils import AnchorGenerator
         >>>
         >>> # load a pre-trained model for classification and return
         >>> # only the features

torchvision/models/detection/mask_rcnn.py

@@ -107,7 +107,7 @@ class MaskRCNN(FasterRCNN):
         >>> import torch
         >>> import torchvision
         >>> from torchvision.models.detection import MaskRCNN
-        >>> from torchvision.models.detection.rpn import AnchorGenerator
+        >>> from torchvision.models.detection.anchor_utils import AnchorGenerator
         >>>
         >>> # load a pre-trained model for classification and return
         >>> # only the features