HTC and SCNet

configs/_base_/datasets/coco_instance_semantic.py

@@ -1,54 +1,65 @@
 # dataset settings
 dataset_type = 'CocoDataset'
 data_root = 'data/coco/'
-img_norm_cfg = dict(
-    mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
+
+# file_client_args = dict(
+#     backend='petrel',
+#     path_mapping=dict({
+#         './data/': 's3://openmmlab/datasets/detection/',
+#         'data/': 's3://openmmlab/datasets/detection/'
+#     }))
+file_client_args = dict(backend='disk')
+
 train_pipeline = [
-    dict(type='LoadImageFromFile'),
+    dict(type='LoadImageFromFile', file_client_args=file_client_args),
     dict(
         type='LoadAnnotations', with_bbox=True, with_mask=True, with_seg=True),
-    dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
-    dict(type='RandomFlip', flip_ratio=0.5),
-    dict(type='Normalize', **img_norm_cfg),
-    dict(type='Pad', size_divisor=32),
-    dict(type='SegRescale', scale_factor=1 / 8),
-    dict(type='DefaultFormatBundle'),
-    dict(
-        type='Collect',
-        keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks', 'gt_semantic_seg']),
+    dict(type='Resize', scale=(1333, 800), keep_ratio=True),
+    dict(type='RandomFlip', prob=0.5),
+    dict(type='PackDetInputs')
 ]
 test_pipeline = [
-    dict(type='LoadImageFromFile'),
+    dict(type='LoadImageFromFile', file_client_args=file_client_args),
+    dict(type='Resize', scale=(1333, 800), keep_ratio=True),
     dict(
-        type='MultiScaleFlipAug',
-        img_scale=(1333, 800),
-        flip=False,
-        transforms=[
-            dict(type='Resize', keep_ratio=True),
-            dict(type='RandomFlip', flip_ratio=0.5),
-            dict(type='Normalize', **img_norm_cfg),
-            dict(type='Pad', size_divisor=32),
-            dict(type='ImageToTensor', keys=['img']),
-            dict(type='Collect', keys=['img']),
-        ])
+        type='PackDetInputs',
+        meta_keys=('img_id', 'img_path', 'ori_shape', 'img_shape',
+                   'scale_factor'))
 ]
-data = dict(
-    samples_per_gpu=2,
-    workers_per_gpu=2,
-    train=dict(
-        type=dataset_type,
-        ann_file=data_root + 'annotations/instances_train2017.json',
-        img_prefix=data_root + 'train2017/',
-        seg_prefix=data_root + 'stuffthingmaps/train2017/',
-        pipeline=train_pipeline),
-    val=dict(
+
+train_dataloader = dict(
+    batch_size=2,
+    num_workers=2,
+    persistent_workers=True,
+    sampler=dict(type='DefaultSampler', shuffle=True),
+    batch_sampler=dict(type='AspectRatioBatchSampler'),
+    dataset=dict(
         type=dataset_type,
-        ann_file=data_root + 'annotations/instances_val2017.json',
-        img_prefix=data_root + 'val2017/',
-        pipeline=test_pipeline),
-    test=dict(
+        data_root=data_root,
+        ann_file='annotations/instances_train2017.json',
+        data_prefix=dict(img='train2017/', seg='stuffthingmaps/train2017/'),
+        filter_cfg=dict(filter_empty_gt=True, min_size=32),
+        pipeline=train_pipeline))
+
+val_dataloader = dict(
+    batch_size=1,
+    num_workers=2,
+    persistent_workers=True,
+    drop_last=False,
+    sampler=dict(type='DefaultSampler', shuffle=False),
+    dataset=dict(
         type=dataset_type,
-        ann_file=data_root + 'annotations/instances_val2017.json',
-        img_prefix=data_root + 'val2017/',
+        data_root=data_root,
+        ann_file='annotations/instances_val2017.json',
+        data_prefix=dict(img='val2017/'),
+        test_mode=True,
         pipeline=test_pipeline))
-evaluation = dict(metric=['bbox', 'segm'])
+
+test_dataloader = val_dataloader
+
+val_evaluator = dict(
+    type='CocoMetric',
+    ann_file=data_root + 'annotations/instances_val2017.json',
+    metric=['bbox', 'segm'],
+    format_only=False)
+test_evaluator = val_evaluator

configs/htc/htc_r101_fpn_20e_coco.py

@@ -1,9 +1,6 @@
-_base_ = './htc_r50_fpn_1x_coco.py'
+_base_ = './htc_r50_fpn_20e_coco.py'
 model = dict(
     backbone=dict(
         depth=101,
         init_cfg=dict(type='Pretrained',
                       checkpoint='torchvision://resnet101')))
-# learning policy
-lr_config = dict(step=[16, 19])
-runner = dict(type='EpochBasedRunner', max_epochs=20)

configs/htc/htc_r50_fpn_1x_coco.py

@@ -1,5 +1,6 @@
 _base_ = './htc_without_semantic_r50_fpn_1x_coco.py'
 model = dict(
+    data_preprocessor=dict(pad_seg=True),
     roi_head=dict(
         semantic_roi_extractor=dict(
             type='SingleRoIExtractor',
@@ -10,47 +11,25 @@
             type='FusedSemanticHead',
             num_ins=5,
             fusion_level=1,
+            seg_scale_factor=1 / 8,
             num_convs=4,
             in_channels=256,
             conv_out_channels=256,
             num_classes=183,
             loss_seg=dict(
                 type='CrossEntropyLoss', ignore_index=255, loss_weight=0.2))))
-data_root = 'data/coco/'
-img_norm_cfg = dict(
-    mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
+
 train_pipeline = [
-    dict(type='LoadImageFromFile'),
     dict(
-        type='LoadAnnotations', with_bbox=True, with_mask=True, with_seg=True),
-    dict(type='Resize', img_scale=(1333, 800), keep_ratio=True),
-    dict(type='RandomFlip', flip_ratio=0.5),
-    dict(type='Normalize', **img_norm_cfg),
-    dict(type='Pad', size_divisor=32),
-    dict(type='SegRescale', scale_factor=1 / 8),
-    dict(type='DefaultFormatBundle'),
-    dict(
-        type='Collect',
-        keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks', 'gt_semantic_seg']),
-]
-test_pipeline = [
-    dict(type='LoadImageFromFile'),
+        type='LoadImageFromFile',
+        file_client_args={{_base_.file_client_args}}),
     dict(
-        type='MultiScaleFlipAug',
-        img_scale=(1333, 800),
-        flip=False,
-        transforms=[
-            dict(type='Resize', keep_ratio=True),
-            dict(type='RandomFlip', flip_ratio=0.5),
-            dict(type='Normalize', **img_norm_cfg),
-            dict(type='Pad', size_divisor=32),
-            dict(type='ImageToTensor', keys=['img']),
-            dict(type='Collect', keys=['img']),
-        ])
+        type='LoadAnnotations', with_bbox=True, with_mask=True, with_seg=True),
+    dict(type='Resize', scale=(1333, 800), keep_ratio=True),
+    dict(type='RandomFlip', prob=0.5),
+    dict(type='PackDetInputs')
 ]
-data = dict(
-    train=dict(
-        seg_prefix=data_root + 'stuffthingmaps/train2017/',
-        pipeline=train_pipeline),
-    val=dict(pipeline=test_pipeline),
-    test=dict(pipeline=test_pipeline))
+train_dataloader = dict(
+    dataset=dict(
+        data_prefix=dict(img='train2017/', seg='stuffthingmaps/train2017/'),
+        pipeline=train_pipeline))

configs/htc/htc_r50_fpn_20e_coco.py

@@ -1,4 +1,16 @@
 _base_ = './htc_r50_fpn_1x_coco.py'
+
 # learning policy
-lr_config = dict(step=[16, 19])
-runner = dict(type='EpochBasedRunner', max_epochs=20)
+max_epochs = 20
+param_scheduler = [
+    dict(
+        type='LinearLR', start_factor=0.001, by_epoch=False, begin=0, end=500),
+    dict(
+        type='MultiStepLR',
+        begin=0,
+        end=max_epochs,
+        by_epoch=True,
+        milestones=[16, 19],
+        gamma=0.1)
+]
+train_cfg = dict(max_epochs=max_epochs)

configs/htc/htc_without_semantic_r50_fpn_1x_coco.py

@@ -5,6 +5,12 @@
 # model settings
 model = dict(
     type='HybridTaskCascade',
+    data_preprocessor=dict(
+        type='DetDataPreprocessor',
+        mean=[123.675, 116.28, 103.53],
+        std=[58.395, 57.12, 57.375],
+        bgr_to_rgb=True,
+        pad_size_divisor=32),
     backbone=dict(
         type='ResNet',
         depth=50,
@@ -215,22 +221,3 @@
             nms=dict(type='nms', iou_threshold=0.5),
             max_per_img=100,
             mask_thr_binary=0.5)))
-img_norm_cfg = dict(
-    mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
-test_pipeline = [
-    dict(type='LoadImageFromFile'),
-    dict(
-        type='MultiScaleFlipAug',
-        img_scale=(1333, 800),
-        flip=False,
-        transforms=[
-            dict(type='Resize', keep_ratio=True),
-            dict(type='RandomFlip', flip_ratio=0.5),
-            dict(type='Normalize', **img_norm_cfg),
-            dict(type='Pad', size_divisor=32),
-            dict(type='ImageToTensor', keys=['img']),
-            dict(type='Collect', keys=['img']),
-        ])
-]
-data = dict(
-    val=dict(pipeline=test_pipeline), test=dict(pipeline=test_pipeline))

configs/htc/htc_x101_32x4d_fpn_16x1_20e_coco.py

@@ -13,7 +13,20 @@
         style='pytorch',
         init_cfg=dict(
             type='Pretrained', checkpoint='open-mmlab://resnext101_32x4d')))
-data = dict(samples_per_gpu=1, workers_per_gpu=1)
+
+train_dataloader = dict(batch_size=1, num_workers=1)
+
 # learning policy
-lr_config = dict(step=[16, 19])
-runner = dict(type='EpochBasedRunner', max_epochs=20)
+max_epochs = 20
+param_scheduler = [
+    dict(
+        type='LinearLR', start_factor=0.001, by_epoch=False, begin=0, end=500),
+    dict(
+        type='MultiStepLR',
+        begin=0,
+        end=max_epochs,
+        by_epoch=True,
+        milestones=[16, 19],
+        gamma=0.1)
+]
+train_cfg = dict(max_epochs=max_epochs)

configs/htc/htc_x101_64x4d_fpn_16x1_20e_coco.py

@@ -1,19 +1,7 @@
-_base_ = './htc_r50_fpn_1x_coco.py'
+_base_ = './htc_x101_32x4d_fpn_16x1_20e_coco.py'
 model = dict(
     backbone=dict(
         type='ResNeXt',
-        depth=101,
         groups=64,
-        base_width=4,
-        num_stages=4,
-        out_indices=(0, 1, 2, 3),
-        frozen_stages=1,
-        norm_cfg=dict(type='BN', requires_grad=True),
-        norm_eval=True,
-        style='pytorch',
         init_cfg=dict(
             type='Pretrained', checkpoint='open-mmlab://resnext101_64x4d')))
-data = dict(samples_per_gpu=1, workers_per_gpu=1)
-# learning policy
-lr_config = dict(step=[16, 19])
-runner = dict(type='EpochBasedRunner', max_epochs=20)

configs/htc/htc_x101_64x4d_fpn_dconv_c3-c5_mstrain_400_1400_16x1_20e_coco.py

@@ -1,43 +1,20 @@
-_base_ = './htc_r50_fpn_1x_coco.py'
+_base_ = './htc_x101_64x4d_fpn_16x1_20e_coco.py'
+
 model = dict(
     backbone=dict(
-        type='ResNeXt',
-        depth=101,
-        groups=64,
-        base_width=4,
-        num_stages=4,
-        out_indices=(0, 1, 2, 3),
-        frozen_stages=1,
-        norm_cfg=dict(type='BN', requires_grad=True),
-        norm_eval=True,
-        style='pytorch',
         dcn=dict(type='DCN', deform_groups=1, fallback_on_stride=False),
-        stage_with_dcn=(False, True, True, True),
-        init_cfg=dict(
-            type='Pretrained', checkpoint='open-mmlab://resnext101_64x4d')))
+        stage_with_dcn=(False, True, True, True)))
+
 # dataset settings
-img_norm_cfg = dict(
-    mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
 train_pipeline = [
     dict(type='LoadImageFromFile'),
     dict(
         type='LoadAnnotations', with_bbox=True, with_mask=True, with_seg=True),
     dict(
-        type='Resize',
-        img_scale=[(1600, 400), (1600, 1400)],
-        multiscale_mode='range',
+        type='RandomResize',
+        scale=[(1600, 400), (1600, 1400)],
         keep_ratio=True),
-    dict(type='RandomFlip', flip_ratio=0.5),
-    dict(type='Normalize', **img_norm_cfg),
-    dict(type='Pad', size_divisor=32),
-    dict(type='SegRescale', scale_factor=1 / 8),
-    dict(type='DefaultFormatBundle'),
-    dict(
-        type='Collect',
-        keys=['img', 'gt_bboxes', 'gt_labels', 'gt_masks', 'gt_semantic_seg']),
+    dict(type='RandomFlip', prob=0.5),
+    dict(type='PackDetInputs')
 ]
-data = dict(
-    samples_per_gpu=1, workers_per_gpu=1, train=dict(pipeline=train_pipeline))
-# learning policy
-lr_config = dict(step=[16, 19])
-runner = dict(type='EpochBasedRunner', max_epochs=20)
+train_dataloader = dict(dataset=dict(pipeline=train_pipeline))

configs/scnet/scnet_r50_fpn_1x_coco.py

@@ -90,6 +90,7 @@
             type='SCNetSemanticHead',
             num_ins=5,
             fusion_level=1,
+            seg_scale_factor=1 / 8,
             num_convs=4,
             in_channels=256,
             conv_out_channels=256,
@@ -112,6 +113,7 @@
             roi_feat_size=7,
             scale_factor=2)))
 
+# TODO
 # uncomment below code to enable test time augmentations
 # img_norm_cfg = dict(
 #     mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)

configs/scnet/scnet_r50_fpn_20e_coco.py

@@ -1,4 +1,15 @@
 _base_ = './scnet_r50_fpn_1x_coco.py'
 # learning policy
-lr_config = dict(step=[16, 19])
-runner = dict(type='EpochBasedRunner', max_epochs=20)
+max_epochs = 20
+param_scheduler = [
+    dict(
+        type='LinearLR', start_factor=0.001, by_epoch=False, begin=0, end=500),
+    dict(
+        type='MultiStepLR',
+        begin=0,
+        end=max_epochs,
+        by_epoch=True,
+        milestones=[16, 19],
+        gamma=0.1)
+]
+train_cfg = dict(max_epochs=max_epochs)

configs/scnet/scnet_x101_64x4d_fpn_8x1_20e_coco.py

@@ -1,7 +1,7 @@
 _base_ = './scnet_x101_64x4d_fpn_20e_coco.py'
-data = dict(samples_per_gpu=1, workers_per_gpu=1)
-optimizer = dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0001)
+train_dataloader = dict(batch_size=1, num_workers=1)
 
+optim_wrapper = dict(optimizer=dict(lr=0.01))
 # NOTE: `auto_scale_lr` is for automatically scaling LR,
 # USER SHOULD NOT CHANGE ITS VALUES.
 # base_batch_size = (8 GPUs) x (1 samples per GPU)

mmdet/datasets/coco.py

@@ -111,10 +111,15 @@ def parse_data_info(self, raw_data_info: dict) -> Union[dict, List[dict]]:
         data_info = {}
 
         img_path = osp.join(self.data_prefix['img'], img_info['file_name'])
-        seg_map_path = img_info['file_name'].replace('jpg', 'png')
+        if self.data_prefix.get('seg', None):
+            seg_map_path = osp.join(
+                self.data_prefix['seg'],
+                img_info['file_name'].replace('jpg', 'png'))
+        else:
+            seg_map_path = None
         data_info['img_path'] = img_path
         data_info['img_id'] = img_info['img_id']
-        data_info['seg_path'] = seg_map_path
+        data_info['seg_map_path'] = seg_map_path
         data_info['height'] = img_info['height']
         data_info['width'] = img_info['width']