Merge branch 'dev-3.x' into dev-3.x-xjy

configs/faster_rcnn/README.md

@@ -39,7 +39,7 @@ We trained with R-50-FPN pytorch style backbone for 1x schedule.
 | R-50-FPN |     L1Loss     |   4.0    |      21.4      |  37.4  |   [config](https://github.com/open-mmlab/mmdetection/tree/master/configs/faster_rcnn/faster-rcnn_r50_fpn_1x_coco.py)   |            [model](https://download.openmmlab.com/mmdetection/v2.0/faster_rcnn/faster_rcnn_r50_fpn_1x_coco/faster_rcnn_r50_fpn_1x_coco_20200130-047c8118.pth) \| [log](https://download.openmmlab.com/mmdetection/v2.0/faster_rcnn/faster_rcnn_r50_fpn_1x_coco/faster_rcnn_r50_fpn_1x_coco_20200130_204655.log.json)             |
 | R-50-FPN |    IoULoss     |          |                |  37.9  | [config](https://github.com/open-mmlab/mmdetection/tree/master/configs/faster_rcnn/faster-rcnn_r50_fpn_iou_1x_coco.py) | [model](https://download.openmmlab.com/mmdetection/v2.0/faster_rcnn/faster_rcnn_r50_fpn_iou_1x_coco/faster_rcnn_r50_fpn_iou_1x_coco_20200506_095954-938e81f0.pth) \| [log](https://download.openmmlab.com/mmdetection/v2.0/faster_rcnn/faster_rcnn_r50_fpn_iou_1x_coco/faster_rcnn_r50_fpn_iou_1x_coco_20200506_095954.log.json) |
 | R-50-FPN |    GIoULoss    |          |                |  37.6  |   [config](https://github.com/open-mmlab/mmdetection/tree/master/configs/faster_rcnn/faster-rcnn_r50_fpn_1x_coco.py)   |            [model](https://download.openmmlab.com/mmdetection/v2.0/faster_rcnn/faster_rcnn_r50_fpn_1x_coco/faster_rcnn_r50_fpn_giou_1x_coco-0eada910.pth) \| [log](https://download.openmmlab.com/mmdetection/v2.0/faster_rcnn/faster_rcnn_r50_fpn_1x_coco/faster_rcnn_r50_fpn_giou_1x_coco_20200505_161120.log.json)            |
-| R-50-FPN | BoundedIoULoss |          |                |  37.4  |   [config](https://github.com/open-mmlab/mmdetection/tree/master/configs/faster_rcnn/faster-rcnn_r50_fpn_1x_coco.py)   |     [model](https://download.openmmlab.com/mmdetection/v2.0/faster_rcnn/faster_rcnn_r50_fpn_1x_coco/faster_rcnn_r50_fpn_bounded_iou_1x_coco-98ad993b.pth) \| [log](https://download.openmmlab.com/mmdetection/v2.0/faster_rcnn/faster_rcnn_r50_fpn_1x_coco/faster_rcnn_r50_fpn_bounded_iou_1x_coco_20200505_160738.log.json)     |
+| R-50-FPN | BoundedIoULoss |          |                |  37.4  |                                 [config](./faster-rcnn_r50_fpn_bounded-iou_1x_coco.py)                                 |     [model](https://download.openmmlab.com/mmdetection/v2.0/faster_rcnn/faster_rcnn_r50_fpn_1x_coco/faster_rcnn_r50_fpn_bounded_iou_1x_coco-98ad993b.pth) \| [log](https://download.openmmlab.com/mmdetection/v2.0/faster_rcnn/faster_rcnn_r50_fpn_1x_coco/faster_rcnn_r50_fpn_bounded_iou_1x_coco_20200505_160738.log.json)     |
 
 ## Pre-trained Models
 

docs/en/get_started.md

@@ -86,7 +86,7 @@ The downloading will take several seconds or more, depending on your network env
 Option (a). If you install MMDetection from source, just run the following command.
 
 ```shell
-python demo/image_demo.py demo/demo.jpg yolov3_mobilenetv2_8xb24-ms-416-300e_coco.py yolov3_mobilenetv2_320_300e_coco_20210719_215349-d18dff72.pth --device cpu --out-file result.jpg
+python demo/image_demo.py demo/demo.jpg yolov3_mobilenetv2_8xb24-320-300e_coco.py yolov3_mobilenetv2_320_300e_coco_20210719_215349-d18dff72.pth --device cpu --out-file result.jpg
 ```
 
 You will see a new image `result.jpg` on your current folder, where bounding boxes are plotted on cars, benches, etc.
@@ -98,7 +98,7 @@ from mmdet.apis import init_detector, inference_detector
 from mmdet.utils import register_all_modules
 
 register_all_modules()
-config_file = 'yolov3_mobilenetv2_8xb24-ms-416-300e_coco.py'
+config_file = 'yolov3_mobilenetv2_8xb24-320-300e_coco.py'
 checkpoint_file = 'yolov3_mobilenetv2_320_300e_coco_20210719_215349-d18dff72.pth'
 model = init_detector(config_file, checkpoint_file, device='cpu')  # or device='cuda:0'
 inference_detector(model, 'demo/demo.jpg')

docs/zh_cn/get_started.md

@@ -83,7 +83,7 @@ mim download mmdet --config yolov3_mobilenetv2_8xb24-320-300e_coco --dest .
 方案 1. 如果你通过源码安装的 MMDetection，那么直接运行以下命令进行验证：
 
 ```shell
-python demo/image_demo.py demo/demo.jpg yolov3_mobilenetv2_8xb24-ms-416-300e_coco.py yolov3_mobilenetv2_320_300e_coco_20210719_215349-d18dff72.pth --device cpu --out-file result.jpg
+python demo/image_demo.py demo/demo.jpg yolov3_mobilenetv2_8xb24-320-300e_coco.py yolov3_mobilenetv2_320_300e_coco_20210719_215349-d18dff72.pth --device cpu --out-file result.jpg
 ```
 
 你会在当前文件夹中看到一个新的图像`result.jpg`，图像中包含有网络预测的检测框。
@@ -95,7 +95,7 @@ from mmdet.apis import init_detector, inference_detector
 from mmdet.utils import register_all_modules
 
 register_all_modules()
-config_file = 'yolov3_mobilenetv2_8xb24-ms-416-300e_coco.py'
+config_file = 'yolov3_mobilenetv2_8xb24-320-300e_coco.py'
 checkpoint_file = 'yolov3_mobilenetv2_320_300e_coco_20210719_215349-d18dff72.pth'
 model = init_detector(config_file, checkpoint_file, device='cpu')  # or device='cuda:0'
 inference_detector(model, 'demo/demo.jpg')

mmdet/engine/hooks/num_class_check_hook.py

@@ -38,9 +38,9 @@ def _check_head(self, runner: Runner, mode: str) -> None:
                  f'CLASSES = ({CLASSES},)')
             from mmdet.models.roi_heads.mask_heads import FusedSemanticHead
             for name, module in model.named_modules():
-                if hasattr(module, 'num_classes'
-                           ) and name != 'rpn_head' and not isinstance(
-                               module, (VGG, FusedSemanticHead)):
+                if hasattr(module, 'num_classes') and not name.endswith(
+                        'rpn_head') and not isinstance(
+                            module, (VGG, FusedSemanticHead)):
                     assert module.num_classes == len(CLASSES), \
                         (f'The `num_classes` ({module.num_classes}) in '
                          f'{module.__class__.__name__} of '

mmdet/models/detectors/panoptic_two_stage_segmentor.py

@@ -101,7 +101,7 @@ def loss(self, batch_inputs: Tensor,
                 x, rpn_data_samples, proposal_cfg=proposal_cfg)
             # avoid get same name with roi_head loss
             keys = rpn_losses.keys()
-            for key in keys:
+            for key in list(keys):
                 if 'loss' in key and 'rpn' not in key:
                     rpn_losses[f'rpn_{key}'] = rpn_losses.pop(key)
             losses.update(rpn_losses)

mmdet/models/detectors/semi_base.py

@@ -1,12 +1,13 @@
 # Copyright (c) OpenMMLab. All rights reserved.
 import copy
-from collections import Sequence
 from typing import Dict, List, Optional, Tuple, Union
 
 import torch
 import torch.nn as nn
 from torch import Tensor
 
+from mmdet.models.utils import (filter_gt_instances, rename_loss_dict,
+                                reweight_loss_dict)
 from mmdet.registry import MODELS
 from mmdet.structures import SampleList
 from mmdet.structures.bbox import bbox_project
@@ -16,7 +17,7 @@
 
 @MODELS.register_module()
 class SemiBaseDetector(BaseDetector):
-    """Base class for semi-supervsed detectors.
+    """Base class for semi-supervised detectors.
 
     Semi-supervised detectors typically consisting of a teacher model
     updated by exponential moving average and a student model updated
@@ -58,17 +59,6 @@ def freeze(model: nn.Module):
         for param in model.parameters():
             param.requires_grad = False
 
-    @staticmethod
-    def reweight_loss(losses: dict, weight: float) -> dict:
-        """Reweight loss for different branches."""
-        for name, loss in losses.items():
-            if 'loss' in name:
-                if isinstance(loss, Sequence):
-                    losses[name] = [item * weight for item in loss]
-                else:
-                    losses[name] = loss * weight
-        return losses
-
     def loss(self, multi_batch_inputs: Dict[str, Tensor],
              multi_batch_data_samples: Dict[str, SampleList]) -> dict:
         """Calculate losses from multi-branch inputs and data samples.
@@ -113,13 +103,10 @@ def loss_by_gt_instances(self, batch_inputs: Tensor,
         Returns:
             dict: A dictionary of loss components
         """
-        gt_loss = {
-            'sup_' + k: v
-            for k, v in self.reweight_loss(
-                self.student.loss(batch_inputs, batch_data_samples),
-                self.semi_train_cfg.get('sup_weight', 1.)).items()
-        }
-        return gt_loss
+
+        losses = self.student.loss(batch_inputs, batch_data_samples)
+        sup_weight = self.semi_train_cfg.get('sup_weight', 1.)
+        return rename_loss_dict('sup_', reweight_loss_dict(losses, sup_weight))
 
     def loss_by_pseudo_instances(self,
                                  batch_inputs: Tensor,
@@ -141,39 +128,17 @@ def loss_by_pseudo_instances(self,
         Returns:
             dict: A dictionary of loss components
         """
-        for data_samples in batch_data_samples:
-            if data_samples.gt_instances.bboxes.shape[0] > 0:
-                data_samples.gt_instances = data_samples.gt_instances[
-                    data_samples.gt_instances.scores >
-                    self.semi_train_cfg.cls_pseudo_thr]
-
+        batch_data_samples = filter_gt_instances(
+            batch_data_samples, score_thr=self.semi_train_cfg.cls_pseudo_thr)
+        losses = self.student.loss(batch_inputs, batch_data_samples)
         pseudo_instances_num = sum([
             len(data_samples.gt_instances)
             for data_samples in batch_data_samples
         ])
         unsup_weight = self.semi_train_cfg.get(
             'unsup_weight', 1.) if pseudo_instances_num > 0 else 0.
-
-        pseudo_loss = {
-            'unsup_' + k: v
-            for k, v in self.reweight_loss(
-                self.student.loss(batch_inputs, batch_data_samples),
-                unsup_weight).items()
-        }
-        return pseudo_loss
-
-    def filter_pseudo_instances(self,
-                                batch_data_samples: SampleList) -> SampleList:
-        """Filter invalid pseudo instances from teacher model."""
-        for data_samples in batch_data_samples:
-            pseudo_bboxes = data_samples.gt_instances.bboxes
-            if pseudo_bboxes.shape[0] > 0:
-                w = pseudo_bboxes[:, 2] - pseudo_bboxes[:, 0]
-                h = pseudo_bboxes[:, 3] - pseudo_bboxes[:, 1]
-                data_samples.gt_instances = data_samples.gt_instances[
-                    (w > self.semi_train_cfg.min_pseudo_bbox_wh[0])
-                    & (h > self.semi_train_cfg.min_pseudo_bbox_wh[1])]
-        return batch_data_samples
+        return rename_loss_dict('unsup_',
+                                reweight_loss_dict(losses, unsup_weight))
 
     @torch.no_grad()
     def get_pseudo_instances(
@@ -198,7 +163,8 @@ def project_pseudo_instances(self, batch_pseudo_instances: SampleList,
                 data_samples.gt_instances.bboxes,
                 torch.tensor(data_samples.homography_matrix).to(
                     self.data_preprocessor.device), data_samples.img_shape)
-        return self.filter_pseudo_instances(batch_data_samples)
+        wh_thr = self.semi_train_cfg.get('min_pseudo_bbox_wh', (1e-2, 1e-2))
+        return filter_gt_instances(batch_data_samples, wh_thr=wh_thr)
 
     def predict(self, batch_inputs: Tensor,
                 batch_data_samples: SampleList) -> SampleList:

mmdet/models/detectors/soft_teacher.py

@@ -6,6 +6,8 @@
 from mmengine.structures import InstanceData
 from torch import Tensor
 
+from mmdet.models.utils import (filter_gt_instances, rename_loss_dict,
+                                reweight_loss_dict)
 from mmdet.registry import MODELS
 from mmdet.structures import SampleList
 from mmdet.structures.bbox import bbox2roi, bbox_project
@@ -77,12 +79,9 @@ def loss_by_pseudo_instances(self,
             x, rpn_results_list, batch_data_samples, batch_info))
         losses.update(**self.rcnn_reg_loss_by_pseudo_instances(
             x, rpn_results_list, batch_data_samples))
-        pseudo_loss = {
-            'unsup_' + k: v
-            for k, v in self.reweight_loss(
-                losses, self.semi_train_cfg.get('unsup_weight', 1.)).items()
-        }
-        return pseudo_loss
+        unsup_weight = self.semi_train_cfg.get('unsup_weight', 1.)
+        return rename_loss_dict('unsup_',
+                                reweight_loss_dict(losses, unsup_weight))
 
     @torch.no_grad()
     def get_pseudo_instances(
@@ -106,10 +105,10 @@ def get_pseudo_instances(
 
         for data_samples, results in zip(batch_data_samples, results_list):
             data_samples.gt_instances = results
-            if data_samples.gt_instances.bboxes.shape[0] > 0:
-                data_samples.gt_instances = data_samples.gt_instances[
-                    data_samples.gt_instances.scores >
-                    self.semi_train_cfg.pseudo_label_initial_score_thr]
+
+        batch_data_samples = filter_gt_instances(
+            batch_data_samples,
+            score_thr=self.semi_train_cfg.pseudo_label_initial_score_thr)
 
         reg_uncs_list = self.compute_uncertainty_with_aug(
             x, batch_data_samples)
@@ -151,12 +150,8 @@ def rpn_loss_by_pseudo_instances(self, x: Tuple[Tensor],
         """
 
         rpn_data_samples = copy.deepcopy(batch_data_samples)
-        for data_samples in rpn_data_samples:
-            if data_samples.gt_instances.bboxes.shape[0] > 0:
-                data_samples.gt_instances = data_samples.gt_instances[
-                    data_samples.gt_instances.scores >
-                    self.semi_train_cfg.rpn_pseudo_thr]
-
+        rpn_data_samples = filter_gt_instances(
+            rpn_data_samples, score_thr=self.semi_train_cfg.rpn_pseudo_thr)
         proposal_cfg = self.student.train_cfg.get('rpn_proposal',
                                                   self.student.test_cfg.rpn)
         # set cat_id of gt_labels to 0 in RPN
@@ -196,11 +191,8 @@ def rcnn_cls_loss_by_pseudo_instances(self, x: Tuple[Tensor],
         """
         rpn_results_list = copy.deepcopy(unsup_rpn_results_list)
         cls_data_samples = copy.deepcopy(batch_data_samples)
-        for data_samples in cls_data_samples:
-            if data_samples.gt_instances.bboxes.shape[0] > 0:
-                data_samples.gt_instances = data_samples.gt_instances[
-                    data_samples.gt_instances.scores >
-                    self.semi_train_cfg.cls_pseudo_thr]
+        cls_data_samples = filter_gt_instances(
+            cls_data_samples, score_thr=self.semi_train_cfg.cls_pseudo_thr)
 
         outputs = unpack_gt_instances(cls_data_samples)
         batch_gt_instances, batch_gt_instances_ignore, _ = outputs
@@ -212,7 +204,6 @@ def rcnn_cls_loss_by_pseudo_instances(self, x: Tuple[Tensor],
             # rename rpn_results.bboxes to rpn_results.priors
             rpn_results = rpn_results_list[i]
             rpn_results.priors = rpn_results.pop('bboxes')
-
             assign_result = self.student.roi_head.bbox_assigner.assign(
                 rpn_results, batch_gt_instances[i],
                 batch_gt_instances_ignore[i])
@@ -226,19 +217,21 @@ def rcnn_cls_loss_by_pseudo_instances(self, x: Tuple[Tensor],
         selected_bboxes = [res.priors for res in sampling_results]
         rois = bbox2roi(selected_bboxes)
         bbox_results = self.student.roi_head._bbox_forward(x, rois)
+        # cls_reg_targets is a tuple of labels, label_weights,
+        # and bbox_targets, bbox_weights
         cls_reg_targets = self.student.roi_head.bbox_head.get_targets(
             sampling_results, self.student.train_cfg.rcnn)
 
         selected_results_list = []
-        for bboxes, data_samples, homography_matrix, img_shape in zip(
+        for bboxes, data_samples, teacher_matrix, teacher_img_shape in zip(
                 selected_bboxes, batch_data_samples,
                 batch_info['homography_matrix'], batch_info['img_shape']):
-            selected_results_list.append(
-                InstanceData(
-                    bboxes=bbox_project(
-                        bboxes, homography_matrix @ torch.tensor(
-                            data_samples.homography_matrix).inverse().to(
-                                self.data_preprocessor.device), img_shape)))
+            student_matrix = torch.tensor(
+                data_samples.homography_matrix, device=teacher_matrix.device)
+            homography_matrix = teacher_matrix @ student_matrix.inverse()
+            projected_bboxes = bbox_project(bboxes, homography_matrix,
+                                            teacher_img_shape)
+            selected_results_list.append(InstanceData(bboxes=projected_bboxes))
 
         with torch.no_grad():
             results_list = self.teacher.roi_head.predict_bbox(
@@ -249,18 +242,18 @@ def rcnn_cls_loss_by_pseudo_instances(self, x: Tuple[Tensor],
                 rescale=False)
             bg_score = torch.cat(
                 [results.scores[:, -1] for results in results_list])
+            # cls_reg_targets[0] is labels
             neg_inds = cls_reg_targets[
                 0] == self.student.roi_head.bbox_head.num_classes
+            # cls_reg_targets[1] is label_weights
             cls_reg_targets[1][neg_inds] = bg_score[neg_inds].detach()
 
         losses = self.student.roi_head.bbox_head.loss(
-            bbox_results['cls_score'],
-            bbox_results['bbox_pred'],
-            rois,
-            *cls_reg_targets,
-            reduction_override='none')
-        losses['loss_cls'] = losses['loss_cls'].sum() / max(
-            cls_reg_targets[1].sum(), 1.0)
+            bbox_results['cls_score'], bbox_results['bbox_pred'], rois,
+            *cls_reg_targets)
+        # cls_reg_targets[1] is label_weights
+        losses['loss_cls'] = losses['loss_cls'] * len(
+            cls_reg_targets[1]) / max(sum(cls_reg_targets[1]), 1.0)
         return losses
 
     def rcnn_reg_loss_by_pseudo_instances(

mmdet/models/detectors/two_stage.py

@@ -172,7 +172,7 @@ def loss(self, batch_inputs: Tensor,
                 x, rpn_data_samples, proposal_cfg=proposal_cfg)
             # avoid get same name with roi_head loss
             keys = rpn_losses.keys()
-            for key in keys:
+            for key in list(keys):
                 if 'loss' in key and 'rpn' not in key:
                     rpn_losses[f'rpn_{key}'] = rpn_losses.pop(key)
             losses.update(rpn_losses)

mmdet/models/utils/__init__.py

@@ -4,12 +4,13 @@
                               get_topk_from_heatmap, transpose_and_gather_feat)
 from .make_divisible import make_divisible
 from .misc import (cat_boxes, center_of_mass, empty_instances,
-                   filter_scores_and_topk, flip_tensor, generate_coordinate,
-                   get_box_tensor, get_box_wh, images_to_levels,
-                   interpolate_as, levels_to_images, mask2ndarray, multi_apply,
-                   samplelist_boxlist2tensor, scale_boxes, select_single_mlvl,
-                   sigmoid_geometric_mean, stack_boxes, unmap,
-                   unpack_gt_instances)
+                   filter_gt_instances, filter_scores_and_topk, flip_tensor,
+                   generate_coordinate, get_box_tensor, get_box_wh,
+                   images_to_levels, interpolate_as, levels_to_images,
+                   mask2ndarray, multi_apply, rename_loss_dict,
+                   reweight_loss_dict, samplelist_boxlist2tensor, scale_boxes,
+                   select_single_mlvl, sigmoid_geometric_mean, stack_boxes,
+                   unmap, unpack_gt_instances)
 from .panoptic_gt_processing import preprocess_panoptic_gt
 from .point_sample import (get_uncertain_point_coords_with_randomness,
                            get_uncertainty)
@@ -24,5 +25,6 @@
     'generate_coordinate', 'levels_to_images', 'mask2ndarray', 'multi_apply',
     'select_single_mlvl', 'unmap', 'images_to_levels',
     'samplelist_boxlist2tensor', 'cat_boxes', 'stack_boxes', 'scale_boxes',
-    'get_box_tensor', 'get_box_wh'
+    'get_box_tensor', 'get_box_wh', 'filter_gt_instances', 'rename_loss_dict',
+    'reweight_loss_dict'
 ]