[Refactor] Refactor decode_head and segmentors and add preprocess_cfg

open-mmlab · Jun 10, 2022 · dd9c411 · dd9c411
1 parent 19f3953
commit dd9c411
Show file tree

Hide file tree

Showing 351 changed files with 1,918 additions and 703 deletions.
diff --git a/configs/_base_/datasets/ade20k.py b/configs/_base_/datasets/ade20k.py
@@ -1,8 +1,6 @@
 # dataset settings
 dataset_type = 'ADE20KDataset'
 data_root = 'data/ade/ADEChallengeData2016'
-img_norm_cfg = dict(
-    mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
 crop_size = (512, 512)
 train_pipeline = [
     dict(type='LoadImageFromFile'),
@@ -11,7 +9,6 @@
     dict(type='RandomCrop', crop_size=crop_size, cat_max_ratio=0.75),
     dict(type='RandomFlip', prob=0.5),
     dict(type='PhotoMetricDistortion'),
-    dict(type='Pad', size=crop_size),
     dict(type='PackSegInputs')
 ]
 test_pipeline = [

diff --git a/configs/_base_/datasets/ade20k_640x640.py b/configs/_base_/datasets/ade20k_640x640.py
@@ -1,8 +1,6 @@
 # dataset settings
 dataset_type = 'ADE20KDataset'
 data_root = 'data/ade/ADEChallengeData2016'
-img_norm_cfg = dict(
-    mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
 crop_size = (640, 640)
 train_pipeline = [
     dict(type='LoadImageFromFile'),
@@ -11,7 +9,6 @@
     dict(type='RandomCrop', crop_size=crop_size, cat_max_ratio=0.75),
     dict(type='RandomFlip', prob=0.5),
     dict(type='PhotoMetricDistortion'),
-    dict(type='Pad', size=crop_size),
     dict(type='PackSegInputs')
 ]
 test_pipeline = [

diff --git a/configs/_base_/datasets/chase_db1.py b/configs/_base_/datasets/chase_db1.py
@@ -1,8 +1,6 @@
 # dataset settings
 dataset_type = 'ChaseDB1Dataset'
 data_root = 'data/CHASE_DB1'
-img_norm_cfg = dict(
-    mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
 img_scale = (960, 999)
 crop_size = (128, 128)
 train_pipeline = [
@@ -12,7 +10,6 @@
     dict(type='RandomCrop', crop_size=crop_size, cat_max_ratio=0.75),
     dict(type='RandomFlip', prob=0.5),
     dict(type='PhotoMetricDistortion'),
-    dict(type='Pad', size=crop_size),
     dict(type='PackSegInputs')
 ]
 test_pipeline = [

diff --git a/configs/_base_/datasets/cityscapes.py b/configs/_base_/datasets/cityscapes.py
@@ -1,8 +1,6 @@
 # dataset settings
 dataset_type = 'CityscapesDataset'
 data_root = 'data/cityscapes/'
-img_norm_cfg = dict(
-    mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
 crop_size = (512, 1024)
 train_pipeline = [
     dict(type='LoadImageFromFile'),
@@ -11,7 +9,6 @@
     dict(type='RandomCrop', crop_size=crop_size, cat_max_ratio=0.75),
     dict(type='RandomFlip', prob=0.5),
     dict(type='PhotoMetricDistortion'),
-    dict(type='Pad', size=crop_size),
     dict(type='PackSegInputs')
 ]
 test_pipeline = [

diff --git a/configs/_base_/datasets/cityscapes_1024x1024.py b/configs/_base_/datasets/cityscapes_1024x1024.py
@@ -1,6 +1,4 @@
 _base_ = './cityscapes.py'
-img_norm_cfg = dict(
-    mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
 crop_size = (1024, 1024)
 train_pipeline = [
     dict(type='LoadImageFromFile'),
@@ -9,7 +7,6 @@
     dict(type='RandomCrop', crop_size=crop_size, cat_max_ratio=0.75),
     dict(type='RandomFlip', prob=0.5),
     dict(type='PhotoMetricDistortion'),
-    dict(type='Pad', size=crop_size),
     dict(type='PackSegInputs')
 ]
 test_pipeline = [

diff --git a/configs/_base_/datasets/cityscapes_768x768.py b/configs/_base_/datasets/cityscapes_768x768.py
@@ -1,6 +1,4 @@
 _base_ = './cityscapes.py'
-img_norm_cfg = dict(
-    mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
 crop_size = (768, 768)
 train_pipeline = [
     dict(type='LoadImageFromFile'),
@@ -9,7 +7,6 @@
     dict(type='RandomCrop', crop_size=crop_size, cat_max_ratio=0.75),
     dict(type='RandomFlip', prob=0.5),
     dict(type='PhotoMetricDistortion'),
-    dict(type='Pad', size=crop_size),
     dict(type='PackSegInputs')
 ]
 test_pipeline = [

diff --git a/configs/_base_/datasets/cityscapes_769x769.py b/configs/_base_/datasets/cityscapes_769x769.py
@@ -1,6 +1,4 @@
 _base_ = './cityscapes.py'
-img_norm_cfg = dict(
-    mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
 crop_size = (769, 769)
 train_pipeline = [
     dict(type='LoadImageFromFile'),
@@ -9,7 +7,6 @@
     dict(type='RandomCrop', crop_size=crop_size, cat_max_ratio=0.75),
     dict(type='RandomFlip', prob=0.5),
     dict(type='PhotoMetricDistortion'),
-    dict(type='Pad', size=crop_size),
     dict(type='PackSegInputs')
 ]
 test_pipeline = [

diff --git a/configs/_base_/datasets/cityscapes_832x832.py b/configs/_base_/datasets/cityscapes_832x832.py
@@ -1,6 +1,4 @@
 _base_ = './cityscapes.py'
-img_norm_cfg = dict(
-    mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
 crop_size = (832, 832)
 train_pipeline = [
     dict(type='LoadImageFromFile'),
@@ -9,7 +7,6 @@
     dict(type='RandomCrop', crop_size=crop_size, cat_max_ratio=0.75),
     dict(type='RandomFlip', prob=0.5),
     dict(type='PhotoMetricDistortion'),
-    dict(type='Pad', size=crop_size),
     dict(type='PackSegInputs')
 ]
 test_pipeline = [

diff --git a/configs/_base_/datasets/coco-stuff10k.py b/configs/_base_/datasets/coco-stuff10k.py
@@ -1,8 +1,6 @@
 # dataset settings
 dataset_type = 'COCOStuffDataset'
 data_root = 'data/coco_stuff10k'
-img_norm_cfg = dict(
-    mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
 crop_size = (512, 512)
 train_pipeline = [
     dict(type='LoadImageFromFile'),
@@ -11,7 +9,6 @@
     dict(type='RandomCrop', crop_size=crop_size, cat_max_ratio=0.75),
     dict(type='RandomFlip', prob=0.5),
     dict(type='PhotoMetricDistortion'),
-    dict(type='Pad', size=crop_size),
     dict(type='PackSegInputs')
 ]
 test_pipeline = [

diff --git a/configs/_base_/datasets/coco-stuff164k.py b/configs/_base_/datasets/coco-stuff164k.py
@@ -1,8 +1,6 @@
 # dataset settings
 dataset_type = 'COCOStuffDataset'
 data_root = 'data/coco_stuff164k'
-img_norm_cfg = dict(
-    mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
 crop_size = (512, 512)
 train_pipeline = [
     dict(type='LoadImageFromFile'),
@@ -11,7 +9,6 @@
     dict(type='RandomCrop', crop_size=crop_size, cat_max_ratio=0.75),
     dict(type='RandomFlip', prob=0.5),
     dict(type='PhotoMetricDistortion'),
-    dict(type='Pad', size=crop_size),
     dict(type='PackSegInputs')
 ]
 test_pipeline = [

diff --git a/configs/_base_/datasets/drive.py b/configs/_base_/datasets/drive.py
@@ -1,8 +1,6 @@
 # dataset settings
 dataset_type = 'DRIVEDataset'
 data_root = 'data/DRIVE'
-img_norm_cfg = dict(
-    mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
 img_scale = (584, 565)
 crop_size = (64, 64)
 train_pipeline = [
@@ -12,7 +10,6 @@
     dict(type='RandomCrop', crop_size=crop_size, cat_max_ratio=0.75),
     dict(type='RandomFlip', prob=0.5),
     dict(type='PhotoMetricDistortion'),
-    dict(type='Pad', size=crop_size),
     dict(type='PackSegInputs')
 ]
 test_pipeline = [

diff --git a/configs/_base_/datasets/hrf.py b/configs/_base_/datasets/hrf.py
@@ -1,8 +1,6 @@
 # dataset settings
 dataset_type = 'HRFDataset'
 data_root = 'data/HRF'
-img_norm_cfg = dict(
-    mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
 img_scale = (2336, 3504)
 crop_size = (256, 256)
 train_pipeline = [
@@ -12,7 +10,6 @@
     dict(type='RandomCrop', crop_size=crop_size, cat_max_ratio=0.75),
     dict(type='RandomFlip', prob=0.5),
     dict(type='PhotoMetricDistortion'),
-    dict(type='Pad', size=crop_size),
     dict(type='PackSegInputs')
 ]
 test_pipeline = [

diff --git a/configs/_base_/datasets/isaid.py b/configs/_base_/datasets/isaid.py
@@ -1,9 +1,6 @@
 # dataset settings
 dataset_type = 'iSAIDDataset'
 data_root = 'data/iSAID'
-
-img_norm_cfg = dict(
-    mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
 """
 This crop_size setting is followed by the implementation of
 `PointFlow: Flowing Semantics Through Points for Aerial Image
@@ -19,7 +16,6 @@
     dict(type='RandomCrop', crop_size=crop_size, cat_max_ratio=0.75),
     dict(type='RandomFlip', prob=0.5),
     dict(type='PhotoMetricDistortion'),
-    dict(type='Pad', size=crop_size),
     dict(type='PackSegInputs')
 ]
 test_pipeline = [

diff --git a/configs/_base_/datasets/loveda.py b/configs/_base_/datasets/loveda.py
@@ -1,8 +1,6 @@
 # dataset settings
 dataset_type = 'LoveDADataset'
 data_root = 'data/loveDA'
-img_norm_cfg = dict(
-    mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
 crop_size = (512, 512)
 train_pipeline = [
     dict(type='LoadImageFromFile'),
@@ -11,7 +9,6 @@
     dict(type='RandomCrop', crop_size=crop_size, cat_max_ratio=0.75),
     dict(type='RandomFlip', prob=0.5),
     dict(type='PhotoMetricDistortion'),
-    dict(type='Pad', size=crop_size),
     dict(type='PackSegInputs')
 ]
 test_pipeline = [

diff --git a/configs/_base_/datasets/pascal_context.py b/configs/_base_/datasets/pascal_context.py
@@ -1,8 +1,6 @@
 # dataset settings
 dataset_type = 'PascalContextDataset'
 data_root = 'data/VOCdevkit/VOC2010/'
-img_norm_cfg = dict(
-    mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
 
 img_scale = (520, 520)
 crop_size = (480, 480)
@@ -14,7 +12,6 @@
     dict(type='RandomCrop', crop_size=crop_size, cat_max_ratio=0.75),
     dict(type='RandomFlip', prob=0.5),
     dict(type='PhotoMetricDistortion'),
-    dict(type='Pad', size=crop_size),
     dict(type='PackSegInputs')
 ]
 test_pipeline = [

diff --git a/configs/_base_/datasets/pascal_context_59.py b/configs/_base_/datasets/pascal_context_59.py
@@ -1,8 +1,6 @@
 # dataset settings
 dataset_type = 'PascalContextDataset59'
 data_root = 'data/VOCdevkit/VOC2010/'
-img_norm_cfg = dict(
-    mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
 
 img_scale = (520, 520)
 crop_size = (480, 480)
@@ -14,7 +12,6 @@
     dict(type='RandomCrop', crop_size=crop_size, cat_max_ratio=0.75),
     dict(type='RandomFlip', prob=0.5),
     dict(type='PhotoMetricDistortion'),
-    dict(type='Pad', size=crop_size),
     dict(type='PackSegInputs')
 ]
 test_pipeline = [

diff --git a/configs/_base_/datasets/pascal_voc12.py b/configs/_base_/datasets/pascal_voc12.py
@@ -1,8 +1,6 @@
 # dataset settings
 dataset_type = 'PascalVOCDataset'
 data_root = 'data/VOCdevkit/VOC2012'
-img_norm_cfg = dict(
-    mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
 crop_size = (512, 512)
 train_pipeline = [
     dict(type='LoadImageFromFile'),
@@ -11,7 +9,6 @@
     dict(type='RandomCrop', crop_size=crop_size, cat_max_ratio=0.75),
     dict(type='RandomFlip', prob=0.5),
     dict(type='PhotoMetricDistortion'),
-    dict(type='Pad', size=crop_size),
     dict(type='PackSegInputs')
 ]
 test_pipeline = [

diff --git a/configs/_base_/datasets/potsdam.py b/configs/_base_/datasets/potsdam.py
@@ -1,8 +1,6 @@
 # dataset settings
 dataset_type = 'PotsdamDataset'
 data_root = 'data/potsdam'
-img_norm_cfg = dict(
-    mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
 crop_size = (512, 512)
 train_pipeline = [
     dict(type='LoadImageFromFile'),
@@ -11,7 +9,6 @@
     dict(type='RandomCrop', crop_size=crop_size, cat_max_ratio=0.75),
     dict(type='RandomFlip', prob=0.5),
     dict(type='PhotoMetricDistortion'),
-    dict(type='Pad', size=crop_size),
     dict(type='PackSegInputs')
 ]
 test_pipeline = [

diff --git a/configs/_base_/datasets/stare.py b/configs/_base_/datasets/stare.py
@@ -1,8 +1,6 @@
 # dataset settings
 dataset_type = 'STAREDataset'
 data_root = 'data/STARE'
-img_norm_cfg = dict(
-    mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
 img_scale = (605, 700)
 crop_size = (128, 128)
 train_pipeline = [
@@ -12,7 +10,6 @@
     dict(type='RandomCrop', crop_size=crop_size, cat_max_ratio=0.75),
     dict(type='RandomFlip', prob=0.5),
     dict(type='PhotoMetricDistortion'),
-    dict(type='Pad', size=crop_size),
     dict(type='PackSegInputs')
 ]
 test_pipeline = [

diff --git a/configs/_base_/datasets/vaihingen.py b/configs/_base_/datasets/vaihingen.py
@@ -1,8 +1,6 @@
 # dataset settings
 dataset_type = 'ISPRSDataset'
 data_root = 'data/vaihingen'
-img_norm_cfg = dict(
-    mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
 crop_size = (512, 512)
 train_pipeline = [
     dict(type='LoadImageFromFile'),
@@ -11,7 +9,6 @@
     dict(type='RandomCrop', crop_size=crop_size, cat_max_ratio=0.75),
     dict(type='RandomFlip', prob=0.5),
     dict(type='PhotoMetricDistortion'),
-    dict(type='Pad', size=crop_size),
     dict(type='PackSegInputs')
 ]
 test_pipeline = [

diff --git a/configs/_base_/models/ann_r50-d8.py b/configs/_base_/models/ann_r50-d8.py
@@ -1,7 +1,14 @@
 # model settings
 norm_cfg = dict(type='SyncBN', requires_grad=True)
+preprocess_cfg = dict(
+    mean=[123.675, 116.28, 103.53],
+    std=[58.395, 57.12, 57.375],
+    to_rgb=True,
+    pad_val=0,
+    seg_pad_val=255)
 model = dict(
     type='EncoderDecoder',
+    preprocess_cfg=preprocess_cfg,
     pretrained='open-mmlab://resnet50_v1c',
     backbone=dict(
         type='ResNetV1c',

diff --git a/configs/_base_/models/apcnet_r50-d8.py b/configs/_base_/models/apcnet_r50-d8.py
@@ -1,7 +1,14 @@
 # model settings
 norm_cfg = dict(type='SyncBN', requires_grad=True)
+preprocess_cfg = dict(
+    mean=[123.675, 116.28, 103.53],
+    std=[58.395, 57.12, 57.375],
+    to_rgb=True,
+    pad_val=0,
+    seg_pad_val=255)
 model = dict(
     type='EncoderDecoder',
+    preprocess_cfg=preprocess_cfg,
     pretrained='open-mmlab://resnet50_v1c',
     backbone=dict(
         type='ResNetV1c',

diff --git a/configs/_base_/models/bisenetv1_r18-d32.py b/configs/_base_/models/bisenetv1_r18-d32.py
@@ -1,7 +1,14 @@
 # model settings
 norm_cfg = dict(type='SyncBN', requires_grad=True)
+preprocess_cfg = dict(
+    mean=[123.675, 116.28, 103.53],
+    std=[58.395, 57.12, 57.375],
+    to_rgb=True,
+    pad_val=0,
+    seg_pad_val=255)
 model = dict(
     type='EncoderDecoder',
+    preprocess_cfg=preprocess_cfg,
     backbone=dict(
         type='BiSeNetV1',
         in_channels=3,

diff --git a/configs/_base_/models/bisenetv2.py b/configs/_base_/models/bisenetv2.py
@@ -1,7 +1,14 @@
 # model settings
 norm_cfg = dict(type='SyncBN', requires_grad=True)
+preprocess_cfg = dict(
+    mean=[123.675, 116.28, 103.53],
+    std=[58.395, 57.12, 57.375],
+    to_rgb=True,
+    pad_val=0,
+    seg_pad_val=255)
 model = dict(
     type='EncoderDecoder',
+    preprocess_cfg=preprocess_cfg,
     pretrained=None,
     backbone=dict(
         type='BiSeNetV2',