Merge branch 'develop' into farseg_in_chns

docs/intro/model_zoo.md

@@ -10,19 +10,20 @@ PaddleRS目前已支持的全部模型如下（标注\*的为遥感专用模型
 |--------|---------|------|
 | 变化检测 | \*BIT | 是 |
 | 变化检测 | \*CDNet | 是 |
+| 变化检测 | \*ChangeFormer | 是 |
+| 变化检测 | \*ChangeStar | 否 |
 | 变化检测 | \*DSAMNet | 是 |
 | 变化检测 | \*DSIFN | 否 |
-| 变化检测 | \*SNUNet | 是 |
-| 变化检测 | \*STANet | 是 |
 | 变化检测 | \*FC-EF | 是 |
 | 变化检测 | \*FC-Siam-conc | 是 |
 | 变化检测 | \*FC-Siam-diff | 是 |
-| 变化检测 | \*ChangeStar | 否 |
-| 变化检测 | \*ChangeFormer | 是 |
+| 变化检测 | \*FCCDN | 是 |
+| 变化检测 | \*SNUNet | 是 |
+| 变化检测 | \*STANet | 是 |
+| 场景分类 | CondenseNetV2 | 是 |
 | 场景分类 | HRNet | 是 |
 | 场景分类 | MobileNetV3 | 是 |
 | 场景分类 | ResNet50-vd | 是 |
-| 场景分类 | CondenseNetV2 | 是 |
 | 图像复原 | DRN | 否 |
 | 图像复原 | ESRGAN | 否 |
 | 图像复原 | LESRCNN | 否 |
@@ -32,5 +33,5 @@ PaddleRS目前已支持的全部模型如下（标注\*的为遥感专用模型
 | 目标检测 | PP-YOLOv2 | 是 |
 | 目标检测 | YOLOv3 | 是 |
 | 图像分割 | DeepLab V3+ | 是 |
-| 图像分割 | UNet | 是 |
 | 图像分割 | \*FarSeg | 否 |
+| 图像分割 | UNet | 是 |

docs/intro/transforms.md

@@ -6,26 +6,26 @@ PaddleRS对不同遥感任务需要的数据预处理/数据增强（合称为
 
 | 数据变换算子名 | 用途                                                     | 任务     | ... |
 | -------------------- | ------------------------------------------------- | -------- | ---- |
-| Resize               | 调整输入影像大小。 | 所有任务 | ... |
-| RandomResize         | 随机调整输入影像大小。 | 所有任务 | ... |
-| ResizeByShort        | 调整输入影像大小，保持纵横比不变（根据短边计算缩放系数）。 | 所有任务 | ... |
-| RandomResizeByShort  | 随机调整输入影像大小，保持纵横比不变（根据短边计算缩放系数）。 | 所有任务 | ... |
-| ResizeByLong         | 调整输入影像大小，保持纵横比不变（根据长边计算缩放系数）。 | 所有任务 | ... |
-| RandomHorizontalFlip | 随机水平翻转输入影像。 | 所有任务 | ... |
-| RandomVerticalFlip   | 随机垂直翻转输入影像。 | 所有任务 | ... |
-| Normalize            | 对输入影像应用标准化。 | 所有任务 | ... |
 | CenterCrop           | 对输入影像进行中心裁剪。 | 所有任务 | ... |
-| RandomCrop           | 对输入影像进行随机中心裁剪。 | 所有任务 | ... |
-| RandomScaleAspect    | 裁剪输入影像并重新缩放到原始尺寸。 | 所有任务 | ... |
-| RandomExpand         | 根据随机偏移扩展输入影像。 | 所有任务 | ... |
-| Pad                  | 将输入影像填充到指定的大小。 | 所有任务 | ... |
+| Dehaze               | 对输入图像进行去雾。 | 所有任务 | ... |
 | MixupImage           | 将两幅影像（及对应的目标检测标注）混合在一起作为新的样本。 | 目标检测 | ... |
-| RandomDistort        | 对输入施加随机色彩变换。 | 所有任务 | ... |
+| Normalize            | 对输入影像应用标准化。 | 所有任务 | ... |
+| Pad                  | 将输入影像填充到指定的大小。 | 所有任务 | ... |
 | RandomBlur           | 对输入施加随机模糊。 | 所有任务 | ... |
-| Dehaze               | 对输入图像进行去雾。 | 所有任务 | ... |
+| RandomCrop           | 对输入影像进行随机中心裁剪。 | 所有任务 | ... |
+| RandomDistort        | 对输入施加随机色彩变换。 | 所有任务 | ... |
+| RandomExpand         | 根据随机偏移扩展输入影像。 | 所有任务 | ... |
+| RandomHorizontalFlip | 随机水平翻转输入影像。 | 所有任务 | ... |
+| RandomResize         | 随机调整输入影像大小。 | 所有任务 | ... |
+| RandomResizeByShort  | 随机调整输入影像大小，保持纵横比不变（根据短边计算缩放系数）。 | 所有任务 | ... |
+| RandomScaleAspect    | 裁剪输入影像并重新缩放到原始尺寸。 | 所有任务 | ... |
+| RandomSwap           | 随机交换两个时相的输入影像。 | 变化检测 | ... |
+| RandomVerticalFlip   | 随机竖直翻转输入影像。 | 所有任务 | ... |
 | ReduceDim            | 对输入图像进行波段降维。 | 所有任务 | ... |
+| Resize               | 调整输入影像大小。 | 所有任务 | ... |
+| ResizeByLong         | 调整输入影像大小，保持纵横比不变（根据长边计算缩放系数）。 | 所有任务 | ... |
+| ResizeByShort        | 调整输入影像大小，保持纵横比不变（根据短边计算缩放系数）。 | 所有任务 | ... |
 | SelectBand           | 对输入影像进行波段选择。 | 所有任务 | ... |
-| RandomSwap           | 随机交换两个时相的输入影像。 | 变化检测 | ... |
 | ...                  | ... | ... | ... |
 
 ## 组合算子

examples/rs_research/config_utils.py

@@ -132,7 +132,7 @@ def parse_args(*args, **kwargs):
         conflict_handler='resolve', parents=[cfg_parser])
     # Global settings
     parser.add_argument('cmd', choices=['train', 'eval'])
-    parser.add_argument('task', choices=['cd', 'clas', 'det', 'seg'])
+    parser.add_argument('task', choices=['cd', 'clas', 'det', 'res', 'seg'])
 
     # Data
     parser.add_argument('--datasets', type=dict, default={})

paddlers/deploy/predictor.py

@@ -105,7 +105,7 @@ class Predictor(object):
                     logging.warning(
                         "Semantic segmentation models do not support TensorRT acceleration, "
                         "TensorRT is forcibly disabled.")
-                elif 'RCNN' in self._model.__class__.__name__:
+                elif self._model.model_type == 'detector' and 'RCNN' in self._model.__class__.__name__:
                     logging.warning(
                         "RCNN models do not support TensorRT acceleration, "
                         "TensorRT is forcibly disabled.")

paddlers/tasks/base.py

@@ -368,7 +368,7 @@ class BaseModel(metaclass=ModelMeta):
                 else:
                     outputs = self.train_step(step, data, self.net)
 
-                scheduler_step(self.optimizer)
+                scheduler_step(self.optimizer, outputs['loss'])
 
                 train_avg_metrics.update(outputs)
                 lr = self.optimizer.get_lr()

paddlers/tasks/object_detector.py

@@ -255,32 +255,18 @@ class BaseDetector(BaseModel):
         """
 
         args = self._pre_train(locals())
+        args.pop('self')
         return self._real_train(**args)
 
     def _pre_train(self, in_args):
         return in_args
 
-    def _real_train(self,
-                    num_epochs,
-                    train_dataset,
-                    train_batch_size=64,
-                    eval_dataset=None,
-                    optimizer=None,
-                    save_interval_epochs=1,
-                    log_interval_steps=10,
-                    save_dir='output',
-                    pretrain_weights='IMAGENET',
-                    learning_rate=.001,
-                    warmup_steps=0,
-                    warmup_start_lr=0.0,
-                    lr_decay_epochs=(216, 243),
-                    lr_decay_gamma=0.1,
-                    metric=None,
-                    use_ema=False,
-                    early_stop=False,
-                    early_stop_patience=5,
-                    use_vdl=True,
-                    resume_checkpoint=None):
+    def _real_train(
+            self, num_epochs, train_dataset, train_batch_size, eval_dataset,
+            optimizer, save_interval_epochs, log_interval_steps, save_dir,
+            pretrain_weights, learning_rate, warmup_steps, warmup_start_lr,
+            lr_decay_epochs, lr_decay_gamma, metric, use_ema, early_stop,
+            early_stop_patience, use_vdl, resume_checkpoint):
 
         if self.status == 'Infer':
             logging.error(

paddlers/tasks/restorer.py

@@ -668,7 +668,7 @@ class DRN(BaseRestorer):
     def __init__(self,
                  losses=None,
                  sr_factor=4,
-                 scale=(2, 4),
+                 scales=(2, 4),
                  n_blocks=30,
                  n_feats=16,
                  n_colors=3,
@@ -677,10 +677,10 @@ class DRN(BaseRestorer):
                  lq_loss_weight=0.1,
                  dual_loss_weight=0.1,
                  **params):
-        if sr_factor != max(scale):
-            raise ValueError(f"`sr_factor` must be equal to `max(scale)`.")
+        if sr_factor != max(scales):
+            raise ValueError(f"`sr_factor` must be equal to `max(scales)`.")
         params.update({
-            'scale': scale,
+            'scale': scales,
             'n_blocks': n_blocks,
             'n_feats': n_feats,
             'n_colors': n_colors,
@@ -689,6 +689,7 @@ class DRN(BaseRestorer):
         })
         self.lq_loss_weight = lq_loss_weight
         self.dual_loss_weight = dual_loss_weight
+        self.scales = scales
         super(DRN, self).__init__(
             model_name='DRN', losses=losses, sr_factor=sr_factor, **params)
 
@@ -720,7 +721,7 @@ class DRN(BaseRestorer):
             lr.extend([
                 F.interpolate(
                     inputs[0], scale_factor=s, mode='bicubic')
-                for s in net.generator.scale[:-1]
+                for s in self.scales[:-1]
             ])
             loss = self.losses(sr[-1], inputs[1])
             for i in range(1, len(sr)):
@@ -733,7 +734,7 @@ class DRN(BaseRestorer):
         elif gan_mode == 'forward_dual':
             sr, lr = inputs[0], inputs[1]
             sr2lr = []
-            n_scales = len(net.generator.scale)
+            n_scales = len(self.scales)
             for i in range(n_scales):
                 sr2lr_i = net.generators[1 + i](sr[i - n_scales])
                 sr2lr.append(sr2lr_i)
@@ -758,6 +759,7 @@ class DRN(BaseRestorer):
         (outputs['loss_prim'] + outputs['loss_dual']).backward()
         self.optimizer.step()
         return {
+            'loss': outputs['loss_prim'] + outputs['loss_dual'],
             'loss_prim': outputs['loss_prim'],
             'loss_dual': outputs['loss_dual']
         }

paddlers/utils/utils.py

@@ -243,20 +243,17 @@ class Timer(Times):
 def to_data_parallel(layers, *args, **kwargs):
     from paddlers.tasks.utils.res_adapters import GANAdapter
     if isinstance(layers, GANAdapter):
-        # Inplace modification for efficiency
-        layers.generators = [
-            paddle.DataParallel(g, *args, **kwargs) for g in layers.generators
-        ]
-        layers.discriminators = [
-            paddle.DataParallel(d, *args, **kwargs)
-            for d in layers.discriminators
-        ]
+        layers = GANAdapter(
+            [to_data_parallel(g, *args, **kwargs) for g in layers.generators], [
+                to_data_parallel(d, *args, **kwargs)
+                for d in layers.discriminators
+            ])
     else:
         layers = paddle.DataParallel(layers, *args, **kwargs)
     return layers
 
 
-def scheduler_step(optimizer):
+def scheduler_step(optimizer, loss=None):
     from paddlers.tasks.utils.res_adapters import OptimizerAdapter
     if not isinstance(optimizer, OptimizerAdapter):
         optimizer = [optimizer]

test_tipc/README.md

@@ -23,11 +23,29 @@
 | 任务类别 | 模型名称 | 基础<br>训练预测 | 更多<br>训练方式 | 更多<br>部署方式 | Slim<br>训练部署 |  更多<br>训练环境  |
 | :--- | :--- |  :----:  | :--------: |  :----:  |   :----:  |   :----:  |
 | 变化检测 | BIT | 支持 | - | - | - |
+| 变化检测 | CDNet | 支持 | - | - | - |
+| 变化检测 | DSAMNet | 支持 | - | - | - |
+| 变化检测 | DSIFN | 支持 | - | - | - |
+| 变化检测 | SNUNet | 支持 | - | - | - |
+| 变化检测 | STANet | 支持 | - | - | - |
+| 变化检测 | FC-EF | 支持 | - | - | - |
+| 变化检测 | FC-Siam-conc | 支持 | - | - | - |
+| 变化检测 | FC-Siam-diff | 支持 | - | - | - |
+| 变化检测 | ChangeFormer | 支持 | - | - | - |
 | 场景分类 | HRNet | 支持 | - | - | - |
+| 场景分类 | MobileNetV3 | 支持 | - | - | - |
+| 场景分类 | ResNet50-vd | 支持 | - | - | - |
+| 图像复原 | DRN | 支持 | - | - | - |
+| 图像复原 | EARGAN | 支持 | - | - | - |
+| 图像复原 | LESRCNN | 支持 | - | - | - |
+| 目标检测 | Faster R-CNN | 支持 | - | - | - |
 | 目标检测 | PP-YOLO | 支持 | - | - | - |
+| 目标检测 | PP-YOLO Tiny | 支持 | - | - | - |
+| 目标检测 | PP-YOLOv2 | 支持 | - | - | - |
+| 目标检测 | YOLOv3 | 支持 | - | - | - |
+| 图像分割 | DeepLab V3+ | 支持 | - | - | - |
 | 图像分割 | UNet | 支持 | - | - | - |
 
-
 ## 3 测试工具简介
 
 ### 3.1 目录介绍

test_tipc/common_func.sh

@@ -86,12 +86,14 @@ function download_and_unzip_dataset() {
         rm -rf "${ds_path}"
     fi
 
-    wget -nc -P "${ds_dir}" "${url}" --no-check-certificate
+    wget -O "${ds_dir}/${zip_name}" "${url}" --no-check-certificate
     
     # The extracted file/directory must have the same name as the zip file.
-    cd "${ds_dir}" && unzip "${zip_name}" \
-        && mv "${zip_name%.*}" ${ds_name} && cd - \
-        && echo "Successfully downloaded ${zip_name} from ${url}. File saved in ${ds_path}. "
+    cd "${ds_dir}" && unzip "${zip_name}"
+    if [ "${zip_name%.*}" != "${ds_name}" ]; then
+        mv "${zip_name%.*}" "${ds_name}"
+    fi
+    cd -
 }
 
 function parse_extra_args() {

test_tipc/config_utils.py

@@ -118,7 +118,7 @@ def parse_args(*args, **kwargs):
         conflict_handler='resolve', parents=[cfg_parser])
     # Global settings
     parser.add_argument('cmd', choices=['train', 'eval'])
-    parser.add_argument('task', choices=['cd', 'clas', 'det', 'seg'])
+    parser.add_argument('task', choices=['cd', 'clas', 'det', 'res', 'seg'])
 
     # Data
     parser.add_argument('--datasets', type=dict, default={})

test_tipc/configs/cd/_base_/airchange.yaml

@@ -60,7 +60,7 @@ download_path: ./test_tipc/data/
 
 num_epochs: 5
 train_batch_size: 4
-save_interval_epochs: 3
+save_interval_epochs: 5
 log_interval_steps: 50
 save_dir: ./test_tipc/output/cd/
 learning_rate: 0.01

test_tipc/configs/cd/bit/bit_airchange.yaml

@@ -1,4 +1,4 @@
-# Basic configurations of BIT with AirChange dataset
+# Configurations of BIT with AirChange dataset
 
 _base_: ../_base_/airchange.yaml
 

test_tipc/configs/cd/bit/bit_levircd.yaml

@@ -1,4 +1,4 @@
-# Basic configurations of BIT with LEVIR-CD dataset
+# Configurations of BIT with LEVIR-CD dataset
 
 _base_: ../_base_/levircd.yaml
 

test_tipc/configs/cd/bit/train_infer_python.txt

@@ -6,7 +6,7 @@ use_gpu:null|null
 --precision:null
 --num_epochs:lite_train_lite_infer=5|lite_train_whole_infer=5|whole_train_whole_infer=10
 --save_dir:adaptive
---train_batch_size:lite_train_lite_infer=4|lite_train_whole_infer=4|whole_train_whole_infer=4
+--train_batch_size:lite_train_lite_infer=4|lite_train_whole_infer=4|whole_train_whole_infer=8
 --model_path:null
 --config:lite_train_lite_infer=./test_tipc/configs/cd/bit/bit_airchange.yaml|lite_train_whole_infer=./test_tipc/configs/cd/bit/bit_airchange.yaml|whole_train_whole_infer=./test_tipc/configs/cd/bit/bit_levircd.yaml
 train_model_name:best_model

test_tipc/configs/cd/cdnet/cdnet_airchange.yaml

@@ -0,0 +1,8 @@
+# Configurations of CDNet with AirChange dataset
+
+_base_: ../_base_/airchange.yaml
+
+save_dir: ./test_tipc/output/cd/cdnet/
+
+model: !Node
+    type: CDNet

test_tipc/configs/cd/cdnet/cdnet_levircd.yaml

@@ -0,0 +1,8 @@
+# Configurations of cdnet with LEVIR-CD dataset
+
+_base_: ../_base_/levircd.yaml
+
+save_dir: ./test_tipc/output/cd/cdnet/
+
+model: !Node
+    type: CDNet

test_tipc/configs/cd/cdnet/train_infer_python.txt

@@ -0,0 +1,53 @@
+===========================train_params===========================
+model_name:cd:cdnet
+python:python
+gpu_list:0|0,1
+use_gpu:null|null
+--precision:null
+--num_epochs:lite_train_lite_infer=5|lite_train_whole_infer=5|whole_train_whole_infer=10
+--save_dir:adaptive
+--train_batch_size:lite_train_lite_infer=4|lite_train_whole_infer=4|whole_train_whole_infer=8
+--model_path:null
+--config:lite_train_lite_infer=./test_tipc/configs/cd/cdnet/cdnet_airchange.yaml|lite_train_whole_infer=./test_tipc/configs/cd/cdnet/cdnet_airchange.yaml|whole_train_whole_infer=./test_tipc/configs/cd/cdnet/cdnet_levircd.yaml
+train_model_name:best_model
+null:null
+##
+trainer:norm
+norm_train:test_tipc/run_task.py train cd
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params===========================
+eval:null
+null:null
+##
+===========================export_params===========================
+--save_dir:adaptive
+--model_dir:adaptive
+--fixed_input_shape:[-1,3,256,256]
+norm_export:deploy/export/export_model.py
+quant_export:null
+fpgm_export:null
+distill_export:null
+export1:null
+export2:null
+===========================infer_params===========================
+infer_model:null
+infer_export:null
+infer_quant:False
+inference:test_tipc/infer.py
+--device:cpu|gpu
+--enable_mkldnn:True
+--cpu_threads:6
+--batch_size:1
+--use_trt:False
+--precision:fp32
+--model_dir:null
+--config:null
+--save_log_path:null
+--benchmark:True
+--model_name:cdnet
+null:null

test_tipc/configs/cd/changeformer/changeformer_airchange.yaml

@@ -0,0 +1,8 @@
+# Configurations of ChangeFormer with AirChange dataset
+
+_base_: ../_base_/airchange.yaml
+
+save_dir: ./test_tipc/output/cd/changeformer/
+
+model: !Node
+    type: ChangeFormer

test_tipc/configs/cd/changeformer/changeformer_levircd.yaml

@@ -0,0 +1,8 @@
+# Configurations of ChangeFormer with LEVIR-CD dataset
+
+_base_: ../_base_/levircd.yaml
+
+save_dir: ./test_tipc/output/cd/changeformer/
+
+model: !Node
+    type: ChangeFormer

test_tipc/configs/cd/changeformer/train_infer_python.txt

@@ -6,14 +6,14 @@ use_gpu:null|null
 --precision:null
 --num_epochs:lite_train_lite_infer=5|lite_train_whole_infer=5|whole_train_whole_infer=10
 --save_dir:adaptive
---train_batch_size:lite_train_lite_infer=4|lite_train_whole_infer=4|whole_train_whole_infer=4
+--train_batch_size:lite_train_lite_infer=4|lite_train_whole_infer=4|whole_train_whole_infer=8
 --model_path:null
+--config:lite_train_lite_infer=./test_tipc/configs/cd/changeformer/changeformer_airchange.yaml|lite_train_whole_infer=./test_tipc/configs/cd/changeformer/changeformer_airchange.yaml|whole_train_whole_infer=./test_tipc/configs/cd/changeformer/changeformer_levircd.yaml
 train_model_name:best_model
-train_infer_file_list:./test_tipc/data/airchange/:./test_tipc/data/airchange/eval.txt
 null:null
 ##
 trainer:norm
-norm_train:test_tipc/run_task.py train cd --config ./test_tipc/configs/cd/changeformer/changeformer.yaml
+norm_train:test_tipc/run_task.py train cd
 pact_train:null
 fpgm_train:null
 distill_train:null
@@ -27,7 +27,7 @@ null:null
 ===========================export_params===========================
 --save_dir:adaptive
 --model_dir:adaptive
---fixed_input_shape:[1,3,256,256]
+--fixed_input_shape:[-1,3,256,256]
 norm_export:deploy/export/export_model.py
 quant_export:null
 fpgm_export:null
@@ -46,7 +46,7 @@ inference:test_tipc/infer.py
 --use_trt:False
 --precision:fp32
 --model_dir:null
---file_list:null:null
+--config:null
 --save_log_path:null
 --benchmark:True
 --model_name:changeformer

test_tipc/configs/cd/dsamnet/dsamnet_airchange.yaml

@@ -0,0 +1,8 @@
+# Configurations of DSAMNet with AirChange dataset
+
+_base_: ../_base_/airchange.yaml
+
+save_dir: ./test_tipc/output/cd/dsamnet/
+
+model: !Node
+    type: DSAMNet

test_tipc/configs/cd/dsamnet/dsamnet_levircd.yaml

@@ -0,0 +1,8 @@
+# Configurations of DSAMNet with LEVIR-CD dataset
+
+_base_: ../_base_/levircd.yaml
+
+save_dir: ./test_tipc/output/cd/dsamnet/
+
+model: !Node
+    type: DSAMNet

test_tipc/configs/cd/dsamnet/train_infer_python.txt

@@ -0,0 +1,53 @@
+===========================train_params===========================
+model_name:cd:dsamnet
+python:python
+gpu_list:0|0,1
+use_gpu:null|null
+--precision:null
+--num_epochs:lite_train_lite_infer=5|lite_train_whole_infer=5|whole_train_whole_infer=10
+--save_dir:adaptive
+--train_batch_size:lite_train_lite_infer=4|lite_train_whole_infer=4|whole_train_whole_infer=8
+--model_path:null
+--config:lite_train_lite_infer=./test_tipc/configs/cd/dsamnet/dsamnet_airchange.yaml|lite_train_whole_infer=./test_tipc/configs/cd/dsamnet/dsamnet_airchange.yaml|whole_train_whole_infer=./test_tipc/configs/cd/dsamnet/dsamnet_levircd.yaml
+train_model_name:best_model
+null:null
+##
+trainer:norm
+norm_train:test_tipc/run_task.py train cd
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params===========================
+eval:null
+null:null
+##
+===========================export_params===========================
+--save_dir:adaptive
+--model_dir:adaptive
+--fixed_input_shape:[-1,3,256,256]
+norm_export:deploy/export/export_model.py
+quant_export:null
+fpgm_export:null
+distill_export:null
+export1:null
+export2:null
+===========================infer_params===========================
+infer_model:null
+infer_export:null
+infer_quant:False
+inference:test_tipc/infer.py
+--device:cpu|gpu
+--enable_mkldnn:True
+--cpu_threads:6
+--batch_size:1
+--use_trt:False
+--precision:fp32
+--model_dir:null
+--config:null
+--save_log_path:null
+--benchmark:True
+--model_name:dsamnet
+null:null

test_tipc/configs/cd/dsifn/dsifn_airchange.yaml

@@ -0,0 +1,8 @@
+# Configurations of DSIFN with AirChange dataset
+
+_base_: ../_base_/airchange.yaml
+
+save_dir: ./test_tipc/output/cd/dsifn/
+
+model: !Node
+    type: DSIFN

test_tipc/configs/cd/dsifn/dsifn_levircd.yaml

@@ -0,0 +1,8 @@
+# Configurations of DSIFN with LEVIR-CD dataset
+
+_base_: ../_base_/levircd.yaml
+
+save_dir: ./test_tipc/output/cd/dsifn/
+
+model: !Node
+    type: DSIFN

test_tipc/configs/cd/dsifn/train_infer_python.txt

@@ -0,0 +1,53 @@
+===========================train_params===========================
+model_name:cd:dsifn
+python:python
+gpu_list:0|0,1
+use_gpu:null|null
+--precision:null
+--num_epochs:lite_train_lite_infer=5|lite_train_whole_infer=5|whole_train_whole_infer=10
+--save_dir:adaptive
+--train_batch_size:lite_train_lite_infer=4|lite_train_whole_infer=4|whole_train_whole_infer=8
+--model_path:null
+--config:lite_train_lite_infer=./test_tipc/configs/cd/dsifn/dsifn_airchange.yaml|lite_train_whole_infer=./test_tipc/configs/cd/dsifn/dsifn_airchange.yaml|whole_train_whole_infer=./test_tipc/configs/cd/dsifn/dsifn_levircd.yaml
+train_model_name:best_model
+null:null
+##
+trainer:norm
+norm_train:test_tipc/run_task.py train cd
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params===========================
+eval:null
+null:null
+##
+===========================export_params===========================
+--save_dir:adaptive
+--model_dir:adaptive
+--fixed_input_shape:[-1,3,256,256]
+norm_export:deploy/export/export_model.py
+quant_export:null
+fpgm_export:null
+distill_export:null
+export1:null
+export2:null
+===========================infer_params===========================
+infer_model:null
+infer_export:null
+infer_quant:False
+inference:test_tipc/infer.py
+--device:cpu|gpu
+--enable_mkldnn:True
+--cpu_threads:6
+--batch_size:1
+--use_trt:False
+--precision:fp32
+--model_dir:null
+--config:null
+--save_log_path:null
+--benchmark:True
+--model_name:dsifn
+null:null

test_tipc/configs/cd/fc_ef/fc_ef_airchange.yaml

@@ -0,0 +1,8 @@
+# Configurations of FC-EF with AirChange dataset
+
+_base_: ../_base_/airchange.yaml
+
+save_dir: ./test_tipc/output/cd/fc_ef/
+
+model: !Node
+    type: FCEarlyFusion

test_tipc/configs/cd/fc_ef/fc_ef_levircd.yaml

@@ -0,0 +1,8 @@
+# Configurations of FC-EF with LEVIR-CD dataset
+
+_base_: ../_base_/levircd.yaml
+
+save_dir: ./test_tipc/output/cd/fc_ef/
+
+model: !Node
+    type: FCEarlyFusion

test_tipc/configs/cd/fc_ef/train_infer_python.txt

@@ -0,0 +1,53 @@
+===========================train_params===========================
+model_name:cd:fc_ef
+python:python
+gpu_list:0|0,1
+use_gpu:null|null
+--precision:null
+--num_epochs:lite_train_lite_infer=5|lite_train_whole_infer=5|whole_train_whole_infer=20
+--save_dir:adaptive
+--train_batch_size:lite_train_lite_infer=4|lite_train_whole_infer=4|whole_train_whole_infer=8
+--model_path:null
+--config:lite_train_lite_infer=./test_tipc/configs/cd/fc_ef/fc_ef_airchange.yaml|lite_train_whole_infer=./test_tipc/configs/cd/fc_ef/fc_ef_airchange.yaml|whole_train_whole_infer=./test_tipc/configs/cd/fc_ef/fc_ef_levircd.yaml
+train_model_name:best_model
+null:null
+##
+trainer:norm
+norm_train:test_tipc/run_task.py train cd
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params===========================
+eval:null
+null:null
+##
+===========================export_params===========================
+--save_dir:adaptive
+--model_dir:adaptive
+--fixed_input_shape:[-1,3,256,256]
+norm_export:deploy/export/export_model.py
+quant_export:null
+fpgm_export:null
+distill_export:null
+export1:null
+export2:null
+===========================infer_params===========================
+infer_model:null
+infer_export:null
+infer_quant:False
+inference:test_tipc/infer.py
+--device:cpu|gpu
+--enable_mkldnn:True
+--cpu_threads:6
+--batch_size:1
+--use_trt:False
+--precision:fp32
+--model_dir:null
+--config:null
+--save_log_path:null
+--benchmark:True
+--model_name:fc_ef
+null:null

test_tipc/configs/cd/fc_siam_conc/fc_siam_conc_airchange.yaml

@@ -0,0 +1,8 @@
+# Configurations of FC-Siam-conc with AirChange dataset
+
+_base_: ../_base_/airchange.yaml
+
+save_dir: ./test_tipc/output/cd/fc_siam_conc/
+
+model: !Node
+    type: FCSiamConc

test_tipc/configs/cd/fc_siam_conc/fc_siam_conc_levircd.yaml

@@ -0,0 +1,8 @@
+# Configurations of FC-Siam-conc with LEVIR-CD dataset
+
+_base_: ../_base_/levircd.yaml
+
+save_dir: ./test_tipc/output/cd/fc_siam_conc/
+
+model: !Node
+    type: FCSiamConc

test_tipc/configs/cd/fc_siam_conc/train_infer_python.txt

@@ -0,0 +1,53 @@
+===========================train_params===========================
+model_name:cd:fc_siam_conc
+python:python
+gpu_list:0|0,1
+use_gpu:null|null
+--precision:null
+--num_epochs:lite_train_lite_infer=5|lite_train_whole_infer=5|whole_train_whole_infer=20
+--save_dir:adaptive
+--train_batch_size:lite_train_lite_infer=4|lite_train_whole_infer=4|whole_train_whole_infer=8
+--model_path:null
+--config:lite_train_lite_infer=./test_tipc/configs/cd/fc_siam_conc/fc_siam_conc_airchange.yaml|lite_train_whole_infer=./test_tipc/configs/cd/fc_siam_conc/fc_siam_conc_airchange.yaml|whole_train_whole_infer=./test_tipc/configs/cd/fc_siam_conc/fc_siam_conc_levircd.yaml
+train_model_name:best_model
+null:null
+##
+trainer:norm
+norm_train:test_tipc/run_task.py train cd
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params===========================
+eval:null
+null:null
+##
+===========================export_params===========================
+--save_dir:adaptive
+--model_dir:adaptive
+--fixed_input_shape:[-1,3,256,256]
+norm_export:deploy/export/export_model.py
+quant_export:null
+fpgm_export:null
+distill_export:null
+export1:null
+export2:null
+===========================infer_params===========================
+infer_model:null
+infer_export:null
+infer_quant:False
+inference:test_tipc/infer.py
+--device:cpu|gpu
+--enable_mkldnn:True
+--cpu_threads:6
+--batch_size:1
+--use_trt:False
+--precision:fp32
+--model_dir:null
+--config:null
+--save_log_path:null
+--benchmark:True
+--model_name:fc_siam_conc
+null:null

test_tipc/configs/cd/fc_siam_diff/fc_siam_diff_airchange.yaml

@@ -0,0 +1,8 @@
+# Configurations of FC-Siam-diff with AirChange dataset
+
+_base_: ../_base_/airchange.yaml
+
+save_dir: ./test_tipc/output/cd/fc_siam_diff/
+
+model: !Node
+    type: FCSiamDiff

test_tipc/configs/cd/fc_siam_diff/fc_siam_diff_levircd.yaml

@@ -0,0 +1,8 @@
+# Configurations of FC-Siam-diff with LEVIR-CD dataset
+
+_base_: ../_base_/levircd.yaml
+
+save_dir: ./test_tipc/output/cd/fc_siam_diff/
+
+model: !Node
+    type: FCSiamDiff

test_tipc/configs/cd/fc_siam_diff/train_infer_python.txt

@@ -0,0 +1,53 @@
+===========================train_params===========================
+model_name:cd:fc_siam_diff
+python:python
+gpu_list:0|0,1
+use_gpu:null|null
+--precision:null
+--num_epochs:lite_train_lite_infer=5|lite_train_whole_infer=5|whole_train_whole_infer=20
+--save_dir:adaptive
+--train_batch_size:lite_train_lite_infer=4|lite_train_whole_infer=4|whole_train_whole_infer=8
+--model_path:null
+--config:lite_train_lite_infer=./test_tipc/configs/cd/fc_siam_diff/fc_siam_diff_airchange.yaml|lite_train_whole_infer=./test_tipc/configs/cd/fc_siam_diff/fc_siam_diff_airchange.yaml|whole_train_whole_infer=./test_tipc/configs/cd/fc_siam_diff/fc_siam_diff_levircd.yaml
+train_model_name:best_model
+null:null
+##
+trainer:norm
+norm_train:test_tipc/run_task.py train cd
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params===========================
+eval:null
+null:null
+##
+===========================export_params===========================
+--save_dir:adaptive
+--model_dir:adaptive
+--fixed_input_shape:[-1,3,256,256]
+norm_export:deploy/export/export_model.py
+quant_export:null
+fpgm_export:null
+distill_export:null
+export1:null
+export2:null
+===========================infer_params===========================
+infer_model:null
+infer_export:null
+infer_quant:False
+inference:test_tipc/infer.py
+--device:cpu|gpu
+--enable_mkldnn:True
+--cpu_threads:6
+--batch_size:1
+--use_trt:False
+--precision:fp32
+--model_dir:null
+--config:null
+--save_log_path:null
+--benchmark:True
+--model_name:fc_siam_diff
+null:null

test_tipc/configs/cd/fccdn/fccdn_airchange.yaml

@@ -0,0 +1,13 @@
+# Configurations of FCCDN with AirChange dataset
+
+_base_: ../_base_/airchange.yaml
+
+save_dir: ./test_tipc/output/cd/fccdn/
+
+model: !Node
+    type: FCCDN
+
+learning_rate: 0.07
+lr_decay_power: 0.6
+log_interval_steps: 100
+save_interval_epochs: 3

test_tipc/configs/cd/fccdn/fccdn_levircd.yaml

@@ -0,0 +1,8 @@
+# Configurations of FCCDN with LEVIR-CD dataset
+
+_base_: ../_base_/levircd.yaml
+
+save_dir: ./test_tipc/output/cd/fccdn/
+
+model: !Node
+    type: FCCDN

test_tipc/configs/cd/fccdn/train_infer_python.txt

@@ -1,19 +1,19 @@
 ===========================train_params===========================
 model_name:cd:fccdn
 python:python
-gpu_list:0
+gpu_list:0|0,1
 use_gpu:null|null
 --precision:null
---num_epochs:lite_train_lite_infer=15|lite_train_whole_infer=15|whole_train_whole_infer=15
+--num_epochs:lite_train_lite_infer=5|lite_train_whole_infer=5|whole_train_whole_infer=10
 --save_dir:adaptive
---train_batch_size:lite_train_lite_infer=4|lite_train_whole_infer=4|whole_train_whole_infer=4
+--train_batch_size:lite_train_lite_infer=4|lite_train_whole_infer=4|whole_train_whole_infer=8
 --model_path:null
+--config:lite_train_lite_infer=./test_tipc/configs/cd/fccdn/fccdn_airchange.yaml|lite_train_whole_infer=./test_tipc/configs/cd/fccdn/fccdn_airchange.yaml|whole_train_whole_infer=./test_tipc/configs/cd/fccdn/fccdn_levircd.yaml
 train_model_name:best_model
-train_infer_file_list:./test_tipc/data/airchange/:./test_tipc/data/airchange/eval.txt
 null:null
 ##
 trainer:norm
-norm_train:test_tipc/run_task.py train cd --config ./test_tipc/configs/cd/fccdn/fccdn.yaml
+norm_train:test_tipc/run_task.py train cd
 pact_train:null
 fpgm_train:null
 distill_train:null
@@ -27,7 +27,7 @@ null:null
 ===========================export_params===========================
 --save_dir:adaptive
 --model_dir:adaptive
---fixed_input_shape:[1,3,256,256]
+--fixed_input_shape:[-1,3,256,256]
 norm_export:deploy/export/export_model.py
 quant_export:null
 fpgm_export:null
@@ -46,8 +46,8 @@ inference:test_tipc/infer.py
 --use_trt:False
 --precision:fp32
 --model_dir:null
---file_list:null:null
+--config:null
 --save_log_path:null
 --benchmark:True
 --model_name:fccdn
-null:null
+null:null

test_tipc/configs/cd/snunet/snunet_airchange.yaml

@@ -0,0 +1,8 @@
+# Configurations of SNUNet with AirChange dataset
+
+_base_: ../_base_/airchange.yaml
+
+save_dir: ./test_tipc/output/cd/snunet/
+
+model: !Node
+    type: SNUNet

test_tipc/configs/cd/snunet/snunet_levircd.yaml

@@ -0,0 +1,8 @@
+# Configurations of SNUNet with LEVIR-CD dataset
+
+_base_: ../_base_/levircd.yaml
+
+save_dir: ./test_tipc/output/cd/snunet/
+
+model: !Node
+    type: SNUNet

test_tipc/configs/cd/snunet/train_infer_python.txt

@@ -0,0 +1,53 @@
+===========================train_params===========================
+model_name:cd:snunet
+python:python
+gpu_list:0|0,1
+use_gpu:null|null
+--precision:null
+--num_epochs:lite_train_lite_infer=5|lite_train_whole_infer=5|whole_train_whole_infer=10
+--save_dir:adaptive
+--train_batch_size:lite_train_lite_infer=4|lite_train_whole_infer=4|whole_train_whole_infer=8
+--model_path:null
+--config:lite_train_lite_infer=./test_tipc/configs/cd/snunet/snunet_airchange.yaml|lite_train_whole_infer=./test_tipc/configs/cd/snunet/snunet_airchange.yaml|whole_train_whole_infer=./test_tipc/configs/cd/snunet/snunet_levircd.yaml
+train_model_name:best_model
+null:null
+##
+trainer:norm
+norm_train:test_tipc/run_task.py train cd
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params===========================
+eval:null
+null:null
+##
+===========================export_params===========================
+--save_dir:adaptive
+--model_dir:adaptive
+--fixed_input_shape:[-1,3,256,256]
+norm_export:deploy/export/export_model.py
+quant_export:null
+fpgm_export:null
+distill_export:null
+export1:null
+export2:null
+===========================infer_params===========================
+infer_model:null
+infer_export:null
+infer_quant:False
+inference:test_tipc/infer.py
+--device:cpu|gpu
+--enable_mkldnn:True
+--cpu_threads:6
+--batch_size:1
+--use_trt:False
+--precision:fp32
+--model_dir:null
+--config:null
+--save_log_path:null
+--benchmark:True
+--model_name:snunet
+null:null

test_tipc/configs/cd/stanet/stanet_airchange.yaml

@@ -0,0 +1,8 @@
+# Configurations of STANet with AirChange dataset
+
+_base_: ../_base_/airchange.yaml
+
+save_dir: ./test_tipc/output/cd/stanet/
+
+model: !Node
+    type: STANet

test_tipc/configs/cd/stanet/stanet_levircd.yaml

@@ -0,0 +1,8 @@
+# Configurations of STANet with LEVIR-CD dataset
+
+_base_: ../_base_/levircd.yaml
+
+save_dir: ./test_tipc/output/cd/stanet/
+
+model: !Node
+    type: STANet

test_tipc/configs/cd/stanet/train_infer_python.txt

@@ -0,0 +1,53 @@
+===========================train_params===========================
+model_name:cd:stanet
+python:python
+gpu_list:0|0,1
+use_gpu:null|null
+--precision:null
+--num_epochs:lite_train_lite_infer=5|lite_train_whole_infer=5|whole_train_whole_infer=10
+--save_dir:adaptive
+--train_batch_size:lite_train_lite_infer=4|lite_train_whole_infer=4|whole_train_whole_infer=8
+--model_path:null
+--config:lite_train_lite_infer=./test_tipc/configs/cd/stanet/stanet_airchange.yaml|lite_train_whole_infer=./test_tipc/configs/cd/stanet/stanet_airchange.yaml|whole_train_whole_infer=./test_tipc/configs/cd/stanet/stanet_levircd.yaml
+train_model_name:best_model
+null:null
+##
+trainer:norm
+norm_train:test_tipc/run_task.py train cd
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params===========================
+eval:null
+null:null
+##
+===========================export_params===========================
+--save_dir:adaptive
+--model_dir:adaptive
+--fixed_input_shape:[-1,3,256,256]
+norm_export:deploy/export/export_model.py
+quant_export:null
+fpgm_export:null
+distill_export:null
+export1:null
+export2:null
+===========================infer_params===========================
+infer_model:null
+infer_export:null
+infer_quant:False
+inference:test_tipc/infer.py
+--device:cpu|gpu
+--enable_mkldnn:True
+--cpu_threads:6
+--batch_size:1
+--use_trt:False
+--precision:fp32
+--model_dir:null
+--config:null
+--save_log_path:null
+--benchmark:True
+--model_name:stanet
+null:null

test_tipc/configs/clas/_base_/ucmerced.yaml

@@ -62,7 +62,7 @@ download_path: ./test_tipc/data/
 
 num_epochs: 2
 train_batch_size: 16
-save_interval_epochs: 5
+save_interval_epochs: 10
 log_interval_steps: 50
 save_dir: ./test_tipc/output/clas/
 learning_rate: 0.01

test_tipc/configs/clas/hrnet/hrnet.yaml

@@ -6,5 +6,5 @@ save_dir: ./test_tipc/output/clas/hrnet/
 
 model: !Node
     type: HRNet_W18_C
-        args:
-            num_classes: 21
+    args:
+        num_classes: 21

test_tipc/configs/clas/hrnet/hrnet_ucmerced.yaml

@@ -0,0 +1,10 @@
+# Configurations of HRNet with UCMerced dataset
+
+_base_: ../_base_/ucmerced.yaml
+
+save_dir: ./test_tipc/output/clas/hrnet/
+
+model: !Node
+    type: HRNet_W18_C
+    args:
+        num_classes: 21

test_tipc/configs/clas/hrnet/train_infer_python.txt

@@ -8,12 +8,12 @@ use_gpu:null|null
 --save_dir:adaptive
 --train_batch_size:lite_train_lite_infer=16|lite_train_whole_infer=16|whole_train_whole_infer=16
 --model_path:null
+--config:lite_train_lite_infer=./test_tipc/configs/clas/hrnet/hrnet_ucmerced.yaml|lite_train_whole_infer=./test_tipc/configs/clas/hrnet/hrnet_ucmerced.yaml|whole_train_whole_infer=./test_tipc/configs/clas/hrnet/hrnet_ucmerced.yaml
 train_model_name:best_model
-train_infer_file_list:./test_tipc/data/ucmerced/:./test_tipc/data/ucmerced/val.txt
 null:null
 ##
 trainer:norm
-norm_train:test_tipc/run_task.py train clas --config ./test_tipc/configs/clas/hrnet/hrnet.yaml
+norm_train:test_tipc/run_task.py train clas
 pact_train:null
 fpgm_train:null
 distill_train:null
@@ -46,7 +46,7 @@ inference:test_tipc/infer.py
 --use_trt:False
 --precision:fp32
 --model_dir:null
---file_list:null:null
+--config:null
 --save_log_path:null
 --benchmark:True
 --model_name:hrnet

test_tipc/configs/clas/mobilenetv3/mobilenetv3_ucmerced.yaml

@@ -0,0 +1,10 @@
+# Configurations of MobileNetV3 with UCMerced dataset
+
+_base_: ../_base_/ucmerced.yaml
+
+save_dir: ./test_tipc/output/clas/mobilenetv3/
+
+model: !Node
+    type: MobileNetV3_small_x1_0
+    args:
+        num_classes: 21

test_tipc/configs/clas/mobilenetv3/train_infer_python.txt

@@ -0,0 +1,53 @@
+===========================train_params===========================
+model_name:clas:mobilenetv3
+python:python
+gpu_list:0|0,1
+use_gpu:null|null
+--precision:null
+--num_epochs:lite_train_lite_infer=3|lite_train_whole_infer=3|whole_train_whole_infer=10
+--save_dir:adaptive
+--train_batch_size:lite_train_lite_infer=16|lite_train_whole_infer=16|whole_train_whole_infer=16
+--model_path:null
+--config:lite_train_lite_infer=./test_tipc/configs/clas/mobilenetv3/mobilenetv3_ucmerced.yaml|lite_train_whole_infer=./test_tipc/configs/clas/mobilenetv3/mobilenetv3_ucmerced.yaml|whole_train_whole_infer=./test_tipc/configs/clas/mobilenetv3/mobilenetv3_ucmerced.yaml
+train_model_name:best_model
+null:null
+##
+trainer:norm
+norm_train:test_tipc/run_task.py train clas
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params===========================
+eval:null
+null:null
+##
+===========================export_params===========================
+--save_dir:adaptive
+--model_dir:adaptive
+--fixed_input_shape:[-1,3,256,256]
+norm_export:deploy/export/export_model.py
+quant_export:null
+fpgm_export:null
+distill_export:null
+export1:null
+export2:null
+===========================infer_params===========================
+infer_model:null
+infer_export:null
+infer_quant:False
+inference:test_tipc/infer.py
+--device:cpu|gpu
+--enable_mkldnn:True
+--cpu_threads:6
+--batch_size:1
+--use_trt:False
+--precision:fp32
+--model_dir:null
+--config:null
+--save_log_path:null
+--benchmark:True
+--model_name:mobilenetv3
+null:null

test_tipc/configs/clas/resnet50_vd/resnet50_vd_ucmerced.yaml

@@ -0,0 +1,10 @@
+# Configurations of ResNet50-vd with UCMerced dataset
+
+_base_: ../_base_/ucmerced.yaml
+
+save_dir: ./test_tipc/output/clas/resnet50_vd/
+
+model: !Node
+    type: ResNet50_vd
+    args:
+        num_classes: 21

test_tipc/configs/clas/resnet50_vd/train_infer_python.txt

@@ -0,0 +1,53 @@
+===========================train_params===========================
+model_name:clas:resnet50_vd
+python:python
+gpu_list:0|0,1
+use_gpu:null|null
+--precision:null
+--num_epochs:lite_train_lite_infer=3|lite_train_whole_infer=3|whole_train_whole_infer=10
+--save_dir:adaptive
+--train_batch_size:lite_train_lite_infer=16|lite_train_whole_infer=16|whole_train_whole_infer=16
+--model_path:null
+--config:lite_train_lite_infer=./test_tipc/configs/clas/resnet50_vd/resnet50_vd_ucmerced.yaml|lite_train_whole_infer=./test_tipc/configs/clas/resnet50_vd/resnet50_vd_ucmerced.yaml|whole_train_whole_infer=./test_tipc/configs/clas/resnet50_vd/resnet50_vd_ucmerced.yaml
+train_model_name:best_model
+null:null
+##
+trainer:norm
+norm_train:test_tipc/run_task.py train clas
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params===========================
+eval:null
+null:null
+##
+===========================export_params===========================
+--save_dir:adaptive
+--model_dir:adaptive
+--fixed_input_shape:[-1,3,256,256]
+norm_export:deploy/export/export_model.py
+quant_export:null
+fpgm_export:null
+distill_export:null
+export1:null
+export2:null
+===========================infer_params===========================
+infer_model:null
+infer_export:null
+infer_quant:False
+inference:test_tipc/infer.py
+--device:cpu|gpu
+--enable_mkldnn:True
+--cpu_threads:6
+--batch_size:1
+--use_trt:False
+--precision:fp32
+--model_dir:null
+--config:null
+--save_log_path:null
+--benchmark:True
+--model_name:resnet50_vd
+null:null

test_tipc/configs/det/_base_/rsod.yaml

@@ -0,0 +1,72 @@
+# Basic configurations of RSOD dataset
+
+datasets:
+    train: !Node
+        type: VOCDetDataset
+        args: 
+            data_dir: ./test_tipc/data/rsod/
+            file_list: ./test_tipc/data/rsod/train.txt
+            label_list: ./test_tipc/data/rsod/labels.txt
+            shuffle: True
+    eval: !Node
+        type: VOCDetDataset
+        args:
+            data_dir: ./test_tipc/data/rsod/
+            file_list: ./test_tipc/data/rsod/val.txt
+            label_list: ./test_tipc/data/rsod/labels.txt
+            shuffle: False
+transforms:
+    train:
+        - !Node
+          type: DecodeImg
+        - !Node
+          type: RandomDistort
+        - !Node
+          type: RandomExpand
+        - !Node
+          type: RandomCrop
+        - !Node
+          type: RandomHorizontalFlip
+        - !Node
+          type: BatchRandomResize
+          args:
+            target_sizes: [320, 352, 384, 416, 448, 480, 512, 544, 576, 608]
+            interp: RANDOM
+        - !Node
+          type: Normalize
+          args:
+            mean: [0.485, 0.456, 0.406]
+            std: [0.229, 0.224, 0.225]
+        - !Node
+          type: ArrangeDetector
+          args: ['train']
+    eval:
+        - !Node
+          type: DecodeImg
+        - !Node
+          type: Resize
+          args:
+            target_size: 608
+            interp: CUBIC
+        - !Node
+          type: Normalize
+          args:
+            mean: [0.485, 0.456, 0.406]
+            std: [0.229, 0.224, 0.225]
+        - !Node
+          type: ArrangeDetector
+          args: ['eval']
+download_on: False
+
+num_epochs: 10
+train_batch_size: 4
+save_interval_epochs: 10
+log_interval_steps: 4
+save_dir: ./test_tipc/output/det/
+learning_rate: 0.0001
+use_vdl: False
+resume_checkpoint: ''
+train:
+    pretrain_weights: COCO
+    warmup_steps: 0
+    warmup_start_lr: 0.0

test_tipc/configs/det/_base_/sarship.yaml

@@ -62,10 +62,10 @@ download_path: ./test_tipc/data/
 
 num_epochs: 10
 train_batch_size: 4
-save_interval_epochs: 5
+save_interval_epochs: 10
 log_interval_steps: 4
 save_dir: ./test_tipc/output/det/
-learning_rate: 0.0005
+learning_rate: 0.0001
 use_vdl: False
 resume_checkpoint: ''
 train:

test_tipc/configs/det/faster_rcnn/faster_rcnn_rsod.yaml

@@ -0,0 +1,10 @@
+# Configurations of Faster R-CNN with RSOD dataset
+
+_base_: ../_base_/rsod.yaml
+
+save_dir: ./test_tipc/output/det/faster_rcnn/
+
+model: !Node
+    type: FasterRCNN
+    args:
+        num_classes: 4

test_tipc/configs/det/faster_rcnn/faster_rcnn_sarship.yaml

@@ -0,0 +1,10 @@
+# Configurations of Faster R-CNN with SARShip dataset
+
+_base_: ../_base_/sarship.yaml
+
+save_dir: ./test_tipc/output/det/faster_rcnn/
+
+model: !Node
+    type: FasterRCNN
+    args:
+        num_classes: 1

test_tipc/configs/det/faster_rcnn/train_infer_python.txt

@@ -0,0 +1,53 @@
+===========================train_params===========================
+model_name:det:faster_rcnn
+python:python
+gpu_list:0|0,1
+use_gpu:null|null
+--precision:null
+--num_epochs:lite_train_lite_infer=3|lite_train_whole_infer=3|whole_train_whole_infer=10
+--save_dir:adaptive
+--train_batch_size:lite_train_lite_infer=4|lite_train_whole_infer=4|whole_train_whole_infer=4
+--model_path:null
+--config:lite_train_lite_infer=./test_tipc/configs/det/faster_rcnn/faster_rcnn_sarship.yaml|lite_train_whole_infer=./test_tipc/configs/det/faster_rcnn/faster_rcnn_sarship.yaml|whole_train_whole_infer=./test_tipc/configs/det/faster_rcnn/faster_rcnn_rsod.yaml
+train_model_name:best_model
+null:null
+##
+trainer:norm
+norm_train:test_tipc/run_task.py train det
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params===========================
+eval:null
+null:null
+##
+===========================export_params===========================
+--save_dir:adaptive
+--model_dir:adaptive
+--fixed_input_shape:[-1,3,608,608]
+norm_export:deploy/export/export_model.py
+quant_export:null
+fpgm_export:null
+distill_export:null
+export1:null
+export2:null
+===========================infer_params===========================
+infer_model:null
+infer_export:null
+infer_quant:False
+inference:test_tipc/infer.py
+--device:cpu|gpu
+--enable_mkldnn:True
+--cpu_threads:6
+--batch_size:1
+--use_trt:False
+--precision:fp32
+--model_dir:null
+--config:null
+--save_log_path:null
+--benchmark:True
+--model_name:faster_rcnn
+null:null

test_tipc/configs/det/ppyolo/ppyolo_rsod.yaml

@@ -0,0 +1,10 @@
+# Configurations of PP-YOLO with RSOD dataset
+
+_base_: ../_base_/rsod.yaml
+
+save_dir: ./test_tipc/output/det/ppyolo/
+
+model: !Node
+    type: PPYOLO
+    args:
+        num_classes: 4

test_tipc/configs/det/ppyolo/ppyolo.yaml → test_tipc/configs/det/ppyolo/ppyolo_sarship.yaml

@@ -1,4 +1,4 @@
-# Basic configurations of PP-YOLO
+# Configurations of PP-YOLO with SARShip dataset
 
 _base_: ../_base_/sarship.yaml
 
@@ -6,5 +6,5 @@ save_dir: ./test_tipc/output/det/ppyolo/
 
 model: !Node
     type: PPYOLO
-        args:
-            num_classes: 1
+    args:
+        num_classes: 1

test_tipc/configs/det/ppyolo/train_infer_python.txt

@@ -4,16 +4,16 @@ python:python
 gpu_list:0|0,1
 use_gpu:null|null
 --precision:null
---num_epochs:lite_train_lite_infer=3|lite_train_whole_infer=3|whole_train_whole_infer=10
+--num_epochs:lite_train_lite_infer=3|lite_train_whole_infer=3|whole_train_whole_infer=20
 --save_dir:adaptive
 --train_batch_size:lite_train_lite_infer=4|lite_train_whole_infer=4|whole_train_whole_infer=4
 --model_path:null
+--config:lite_train_lite_infer=./test_tipc/configs/det/ppyolo/ppyolo_sarship.yaml|lite_train_whole_infer=./test_tipc/configs/det/ppyolo/ppyolo_sarship.yaml|whole_train_whole_infer=./test_tipc/configs/det/ppyolo/ppyolo_rsod.yaml
 train_model_name:best_model
-train_infer_file_list:./test_tipc/data/sarship/:./test_tipc/data/sarship/eval.txt
 null:null
 ##
 trainer:norm
-norm_train:test_tipc/run_task.py train det --config ./test_tipc/configs/det/ppyolo/ppyolo.yaml
+norm_train:test_tipc/run_task.py train det
 pact_train:null
 fpgm_train:null
 distill_train:null
@@ -46,7 +46,7 @@ inference:test_tipc/infer.py
 --use_trt:False
 --precision:fp32
 --model_dir:null
---file_list:null:null
+--config:null
 --save_log_path:null
 --benchmark:True
 --model_name:ppyolo

test_tipc/configs/det/ppyolo_tiny/ppyolo_tiny_rsod.yaml

@@ -0,0 +1,10 @@
+# Configurations of PP-YOLO Tiny with RSOD dataset
+
+_base_: ../_base_/rsod.yaml
+
+save_dir: ./test_tipc/output/det/ppyolo_tiny/
+
+model: !Node
+    type: PPYOLOTiny
+    args:
+        num_classes: 4

test_tipc/configs/det/ppyolo_tiny/ppyolo_tiny_sarship.yaml

@@ -0,0 +1,10 @@
+# Configurations of PP-YOLO Tiny with SARShip dataset
+
+_base_: ../_base_/sarship.yaml
+
+save_dir: ./test_tipc/output/det/ppyolo_tiny/
+
+model: !Node
+    type: PPYOLOTiny
+    args:
+        num_classes: 1

test_tipc/configs/det/ppyolo_tiny/train_infer_python.txt

@@ -0,0 +1,53 @@
+===========================train_params===========================
+model_name:det:ppyolo_tiny
+python:python
+gpu_list:0|0,1
+use_gpu:null|null
+--precision:null
+--num_epochs:lite_train_lite_infer=3|lite_train_whole_infer=3|whole_train_whole_infer=20
+--save_dir:adaptive
+--train_batch_size:lite_train_lite_infer=4|lite_train_whole_infer=4|whole_train_whole_infer=4
+--model_path:null
+--config:lite_train_lite_infer=./test_tipc/configs/det/ppyolo_tiny/ppyolo_tiny_sarship.yaml|lite_train_whole_infer=./test_tipc/configs/det/ppyolo_tiny/ppyolo_tiny_sarship.yaml|whole_train_whole_infer=./test_tipc/configs/det/ppyolo_tiny/ppyolo_tiny_rsod.yaml
+train_model_name:best_model
+null:null
+##
+trainer:norm
+norm_train:test_tipc/run_task.py train det
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params===========================
+eval:null
+null:null
+##
+===========================export_params===========================
+--save_dir:adaptive
+--model_dir:adaptive
+--fixed_input_shape:[-1,3,608,608]
+norm_export:deploy/export/export_model.py
+quant_export:null
+fpgm_export:null
+distill_export:null
+export1:null
+export2:null
+===========================infer_params===========================
+infer_model:null
+infer_export:null
+infer_quant:False
+inference:test_tipc/infer.py
+--device:cpu|gpu
+--enable_mkldnn:True
+--cpu_threads:6
+--batch_size:1
+--use_trt:False
+--precision:fp32
+--model_dir:null
+--config:null
+--save_log_path:null
+--benchmark:True
+--model_name:ppyolo_tiny
+null:null

test_tipc/configs/det/ppyolov2/ppyolov2_rsod.yaml

@@ -0,0 +1,10 @@
+# Configurations of PP-YOLOv2 with RSOD dataset
+
+_base_: ../_base_/rsod.yaml
+
+save_dir: ./test_tipc/output/det/ppyolov2/
+
+model: !Node
+    type: PPYOLOv2
+    args:
+        num_classes: 4

test_tipc/configs/det/ppyolov2/ppyolov2_sarship.yaml

@@ -0,0 +1,10 @@
+# Configurations of PP-YOLOv2 with SARShip dataset
+
+_base_: ../_base_/sarship.yaml
+
+save_dir: ./test_tipc/output/det/ppyolov2/
+
+model: !Node
+    type: PPYOLOv2
+    args:
+        num_classes: 1

test_tipc/configs/det/ppyolov2/train_infer_python.txt

@@ -0,0 +1,53 @@
+===========================train_params===========================
+model_name:det:ppyolov2
+python:python
+gpu_list:0|0,1
+use_gpu:null|null
+--precision:null
+--num_epochs:lite_train_lite_infer=3|lite_train_whole_infer=3|whole_train_whole_infer=20
+--save_dir:adaptive
+--train_batch_size:lite_train_lite_infer=4|lite_train_whole_infer=4|whole_train_whole_infer=4
+--model_path:null
+--config:lite_train_lite_infer=./test_tipc/configs/det/ppyolov2/ppyolov2_sarship.yaml|lite_train_whole_infer=./test_tipc/configs/det/ppyolov2/ppyolov2_sarship.yaml|whole_train_whole_infer=./test_tipc/configs/det/ppyolov2/ppyolov2_rsod.yaml
+train_model_name:best_model
+null:null
+##
+trainer:norm
+norm_train:test_tipc/run_task.py train det
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params===========================
+eval:null
+null:null
+##
+===========================export_params===========================
+--save_dir:adaptive
+--model_dir:adaptive
+--fixed_input_shape:[-1,3,608,608]
+norm_export:deploy/export/export_model.py
+quant_export:null
+fpgm_export:null
+distill_export:null
+export1:null
+export2:null
+===========================infer_params===========================
+infer_model:null
+infer_export:null
+infer_quant:False
+inference:test_tipc/infer.py
+--device:cpu|gpu
+--enable_mkldnn:True
+--cpu_threads:6
+--batch_size:1
+--use_trt:False
+--precision:fp32
+--model_dir:null
+--config:null
+--save_log_path:null
+--benchmark:True
+--model_name:ppyolov2
+null:null

test_tipc/configs/det/yolov3/train_infer_python.txt

@@ -0,0 +1,53 @@
+===========================train_params===========================
+model_name:det:yolov3
+python:python
+gpu_list:0|0,1
+use_gpu:null|null
+--precision:null
+--num_epochs:lite_train_lite_infer=3|lite_train_whole_infer=3|whole_train_whole_infer=10
+--save_dir:adaptive
+--train_batch_size:lite_train_lite_infer=4|lite_train_whole_infer=4|whole_train_whole_infer=4
+--model_path:null
+--config:lite_train_lite_infer=./test_tipc/configs/det/yolov3/yolov3_sarship.yaml|lite_train_whole_infer=./test_tipc/configs/det/yolov3/yolov3_sarship.yaml|whole_train_whole_infer=./test_tipc/configs/det/yolov3/yolov3_rsod.yaml
+train_model_name:best_model
+null:null
+##
+trainer:norm
+norm_train:test_tipc/run_task.py train det
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params===========================
+eval:null
+null:null
+##
+===========================export_params===========================
+--save_dir:adaptive
+--model_dir:adaptive
+--fixed_input_shape:[-1,3,608,608]
+norm_export:deploy/export/export_model.py
+quant_export:null
+fpgm_export:null
+distill_export:null
+export1:null
+export2:null
+===========================infer_params===========================
+infer_model:null
+infer_export:null
+infer_quant:False
+inference:test_tipc/infer.py
+--device:cpu|gpu
+--enable_mkldnn:True
+--cpu_threads:6
+--batch_size:1
+--use_trt:False
+--precision:fp32
+--model_dir:null
+--config:null
+--save_log_path:null
+--benchmark:True
+--model_name:yolov3
+null:null

test_tipc/configs/det/yolov3/yolov3_rsod.yaml

@@ -0,0 +1,10 @@
+# Configurations of YOLOv3 with RSOD dataset
+
+_base_: ../_base_/rsod.yaml
+
+save_dir: ./test_tipc/output/det/yolov3/
+
+model: !Node
+    type: YOLOv3
+    args:
+        num_classes: 4

test_tipc/configs/det/yolov3/yolov3_sarship.yaml

@@ -0,0 +1,10 @@
+# Configurations of YOLOv3 with SARShip dataset
+
+_base_: ../_base_/sarship.yaml
+
+save_dir: ./test_tipc/output/det/yolov3/
+
+model: !Node
+    type: YOLOv3
+    args:
+        num_classes: 1

test_tipc/configs/res/_base_/rssr.yaml

@@ -0,0 +1,72 @@
+# Basic configurations of RSSR dataset
+
+datasets:
+    train: !Node
+        type: ResDataset
+        args: 
+            data_dir: ./test_tipc/data/rssr/
+            file_list: ./test_tipc/data/rssr/train.txt
+            num_workers: 0
+            shuffle: True
+            sr_factor: 4
+    eval: !Node
+        type: ResDataset
+        args:
+            data_dir: ./test_tipc/data/rssr/
+            file_list: ./test_tipc/data/rssr/val.txt
+            num_workers: 0
+            shuffle: False
+            sr_factor: 4
+transforms:
+    train:
+        - !Node
+          type: DecodeImg
+        - !Node
+          type: RandomCrop
+          args:
+            crop_size: 32
+        - !Node
+          type: RandomHorizontalFlip
+          args:
+            prob: 0.5
+        - !Node
+          type: RandomVerticalFlip
+          args:
+            prob: 0.5
+        - !Node
+          type: Normalize
+          args:
+            mean: [0.0, 0.0, 0.0]
+            std: [1.0, 1.0, 1.0]
+        - !Node
+          type: ArrangeRestorer
+          args: ['train']
+    eval:
+        - !Node
+          type: DecodeImg
+        - !Node
+          type: Resize
+          args:
+            target_size: 256
+        - !Node
+          type: Normalize
+          args:
+            mean: [0.0, 0.0, 0.0]
+            std: [1.0, 1.0, 1.0]
+        - !Node
+          type: ArrangeRestorer
+          args: ['eval']
+download_on: False
+download_url: https://paddlers.bj.bcebos.com/datasets/rssr.zip
+download_path: ./test_tipc/data/
+
+num_epochs: 10
+train_batch_size: 4
+save_interval_epochs: 10
+log_interval_steps: 10
+save_dir: ./test_tipc/output/res/
+learning_rate: 0.0005
+early_stop: False
+early_stop_patience: 5
+use_vdl: False
+resume_checkpoint: ''

test_tipc/configs/res/drn/drn_rssr.yaml

@@ -0,0 +1,8 @@
+# Configurations of DRN with RSSR dataset
+
+_base_: ../_base_/rssr.yaml
+
+save_dir: ./test_tipc/output/res/drn/
+
+model: !Node
+    type: DRN

test_tipc/configs/res/drn/train_infer_python.txt

@@ -0,0 +1,53 @@
+===========================train_params===========================
+model_name:res:drn
+python:python
+gpu_list:0|0,1
+use_gpu:null|null
+--precision:null
+--num_epochs:lite_train_lite_infer=3|lite_train_whole_infer=3|whole_train_whole_infer=20
+--save_dir:adaptive
+--train_batch_size:lite_train_lite_infer=4|lite_train_whole_infer=4|whole_train_whole_infer=4
+--model_path:null
+--config:lite_train_lite_infer=./test_tipc/configs/res/drn/drn_rssr.yaml|lite_train_whole_infer=./test_tipc/configs/res/drn/drn_rssr.yaml|whole_train_whole_infer=./test_tipc/configs/res/drn/drn_rssr.yaml
+train_model_name:best_model
+null:null
+##
+trainer:norm
+norm_train:test_tipc/run_task.py train res
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params===========================
+eval:null
+null:null
+##
+===========================export_params===========================
+--save_dir:adaptive
+--model_dir:adaptive
+--fixed_input_shape:[-1,3,256,256]
+norm_export:deploy/export/export_model.py
+quant_export:null
+fpgm_export:null
+distill_export:null
+export1:null
+export2:null
+===========================infer_params===========================
+infer_model:null
+infer_export:null
+infer_quant:False
+inference:test_tipc/infer.py
+--device:cpu|gpu
+--enable_mkldnn:True
+--cpu_threads:6
+--batch_size:1
+--use_trt:False
+--precision:fp32
+--model_dir:null
+--config:null
+--save_log_path:null
+--benchmark:True
+--model_name:drn
+null:null

test_tipc/configs/res/esrgan/esrgan_rssr.yaml

@@ -0,0 +1,8 @@
+# Configurations of ESRGAN with RSSR dataset
+
+_base_: ../_base_/rssr.yaml
+
+save_dir: ./test_tipc/output/res/esrgan/
+
+model: !Node
+    type: ESRGAN

test_tipc/configs/res/esrgan/train_infer_python.txt

@@ -0,0 +1,53 @@
+===========================train_params===========================
+model_name:res:esrgan
+python:python
+gpu_list:0|0,1
+use_gpu:null|null
+--precision:null
+--num_epochs:lite_train_lite_infer=3|lite_train_whole_infer=3|whole_train_whole_infer=20
+--save_dir:adaptive
+--train_batch_size:lite_train_lite_infer=4|lite_train_whole_infer=4|whole_train_whole_infer=4
+--model_path:null
+--config:lite_train_lite_infer=./test_tipc/configs/res/esrgan/esrgan_rssr.yaml|lite_train_whole_infer=./test_tipc/configs/res/esrgan/esrgan_rssr.yaml|whole_train_whole_infer=./test_tipc/configs/res/esrgan/esrgan_rssr.yaml
+train_model_name:best_model
+null:null
+##
+trainer:norm
+norm_train:test_tipc/run_task.py train res
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params===========================
+eval:null
+null:null
+##
+===========================export_params===========================
+--save_dir:adaptive
+--model_dir:adaptive
+--fixed_input_shape:[-1,3,256,256]
+norm_export:deploy/export/export_model.py
+quant_export:null
+fpgm_export:null
+distill_export:null
+export1:null
+export2:null
+===========================infer_params===========================
+infer_model:null
+infer_export:null
+infer_quant:False
+inference:test_tipc/infer.py
+--device:cpu|gpu
+--enable_mkldnn:True
+--cpu_threads:6
+--batch_size:1
+--use_trt:False
+--precision:fp32
+--model_dir:null
+--config:null
+--save_log_path:null
+--benchmark:True
+--model_name:esrgan
+null:null

test_tipc/configs/res/lesrcnn/lesrcnn_rssr.yaml

@@ -0,0 +1,8 @@
+# Configurations of LESRCNN with RSSR dataset
+
+_base_: ../_base_/rssr.yaml
+
+save_dir: ./test_tipc/output/res/lesrcnn/
+
+model: !Node
+    type: LESRCNN

test_tipc/configs/res/lesrcnn/train_infer_python.txt

@@ -0,0 +1,53 @@
+===========================train_params===========================
+model_name:res:lesrcnn
+python:python
+gpu_list:0|0,1
+use_gpu:null|null
+--precision:null
+--num_epochs:lite_train_lite_infer=3|lite_train_whole_infer=3|whole_train_whole_infer=20
+--save_dir:adaptive
+--train_batch_size:lite_train_lite_infer=4|lite_train_whole_infer=4|whole_train_whole_infer=4
+--model_path:null
+--config:lite_train_lite_infer=./test_tipc/configs/res/lesrcnn/lesrcnn_rssr.yaml|lite_train_whole_infer=./test_tipc/configs/res/lesrcnn/lesrcnn_rssr.yaml|whole_train_whole_infer=./test_tipc/configs/res/lesrcnn/lesrcnn_rssr.yaml
+train_model_name:best_model
+null:null
+##
+trainer:norm
+norm_train:test_tipc/run_task.py train res
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params===========================
+eval:null
+null:null
+##
+===========================export_params===========================
+--save_dir:adaptive
+--model_dir:adaptive
+--fixed_input_shape:[-1,3,256,256]
+norm_export:deploy/export/export_model.py
+quant_export:null
+fpgm_export:null
+distill_export:null
+export1:null
+export2:null
+===========================infer_params===========================
+infer_model:null
+infer_export:null
+infer_quant:False
+inference:test_tipc/infer.py
+--device:cpu|gpu
+--enable_mkldnn:True
+--cpu_threads:6
+--batch_size:1
+--use_trt:False
+--precision:fp32
+--model_dir:null
+--config:null
+--save_log_path:null
+--benchmark:True
+--model_name:lesrcnn
+null:null

test_tipc/configs/seg/_base_/rsseg.yaml

@@ -58,7 +58,7 @@ download_path: ./test_tipc/data/
 
 num_epochs: 10
 train_batch_size: 4
-save_interval_epochs: 5
+save_interval_epochs: 10
 log_interval_steps: 4
 save_dir: ./test_tipc/output/seg/
 learning_rate: 0.001

test_tipc/configs/seg/deeplabv3p/deeplabv3p_rsseg.yaml

@@ -0,0 +1,11 @@
+# Configurations of DeepLab V3+ with RSSeg dataset
+
+_base_: ../_base_/rsseg.yaml
+
+save_dir: ./test_tipc/output/seg/deeplabv3p/
+
+model: !Node
+    type: DeepLabV3P
+    args:
+        in_channels: 10
+        num_classes: 5

test_tipc/configs/seg/deeplabv3p/train_infer_python.txt

@@ -0,0 +1,53 @@
+===========================train_params===========================
+model_name:seg:deeplabv3p
+python:python
+gpu_list:0|0,1
+use_gpu:null|null
+--precision:null
+--num_epochs:lite_train_lite_infer=3|lite_train_whole_infer=3|whole_train_whole_infer=30
+--save_dir:adaptive
+--train_batch_size:lite_train_lite_infer=4|lite_train_whole_infer=4|whole_train_whole_infer=4
+--model_path:null
+--config:lite_train_lite_infer=./test_tipc/configs/seg/deeplabv3p/deeplabv3p_rsseg.yaml|lite_train_whole_infer=./test_tipc/configs/seg/deeplabv3p/deeplabv3p_rsseg.yaml|whole_train_whole_infer=./test_tipc/configs/seg/deeplabv3p/deeplabv3p_rsseg.yaml
+train_model_name:best_model
+null:null
+##
+trainer:norm
+norm_train:test_tipc/run_task.py train seg
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params===========================
+eval:null
+null:null
+##
+===========================export_params===========================
+--save_dir:adaptive
+--model_dir:adaptive
+--fixed_input_shape:[-1,10,512,512]
+norm_export:deploy/export/export_model.py
+quant_export:null
+fpgm_export:null
+distill_export:null
+export1:null
+export2:null
+===========================infer_params===========================
+infer_model:null
+infer_export:null
+infer_quant:False
+inference:test_tipc/infer.py
+--device:cpu|gpu
+--enable_mkldnn:True
+--cpu_threads:6
+--batch_size:1
+--use_trt:False
+--precision:fp32
+--model_dir:null
+--config:null
+--save_log_path:null
+--benchmark:True
+--model_name:deeplabv3p
+null:null

test_tipc/configs/seg/unet/train_infer_python.txt

@@ -4,16 +4,16 @@ python:python
 gpu_list:0|0,1
 use_gpu:null|null
 --precision:null
---num_epochs:lite_train_lite_infer=3|lite_train_whole_infer=3|whole_train_whole_infer=10
+--num_epochs:lite_train_lite_infer=3|lite_train_whole_infer=3|whole_train_whole_infer=20
 --save_dir:adaptive
 --train_batch_size:lite_train_lite_infer=4|lite_train_whole_infer=4|whole_train_whole_infer=4
 --model_path:null
+--config:lite_train_lite_infer=./test_tipc/configs/seg/unet/unet_rsseg.yaml|lite_train_whole_infer=./test_tipc/configs/seg/unet/unet_rsseg.yaml|whole_train_whole_infer=./test_tipc/configs/seg/unet/unet_rsseg.yaml
 train_model_name:best_model
-train_infer_file_list:./test_tipc/data/rsseg/:./test_tipc/data/rsseg/val.txt
 null:null
 ##
 trainer:norm
-norm_train:test_tipc/run_task.py train seg --config ./test_tipc/configs/seg/unet/unet.yaml
+norm_train:test_tipc/run_task.py train seg
 pact_train:null
 fpgm_train:null
 distill_train:null
@@ -46,7 +46,7 @@ inference:test_tipc/infer.py
 --use_trt:False
 --precision:fp32
 --model_dir:null
---file_list:null:null
+--config:null
 --save_log_path:null
 --benchmark:True
 --model_name:unet

test_tipc/configs/seg/unet/unet.yaml

@@ -1,11 +0,0 @@
-# Basic configurations of UNet
-
-_base_: ../_base_/rsseg.yaml
-
-save_dir: ./test_tipc/output/seg/unet/
-
-model: !Node
-       type: UNet
-       args:
-           in_channels: 10
-           num_classes: 5

test_tipc/configs/seg/unet/unet_rsseg.yaml

@@ -0,0 +1,11 @@
+# Configurations of UNet with RSSeg dataset
+
+_base_: ../_base_/rsseg.yaml
+
+save_dir: ./test_tipc/output/seg/unet/
+
+model: !Node
+    type: UNet
+    args:
+        in_channels: 10
+        num_classes: 5

test_tipc/docs/test_train_inference_python.md

@@ -6,22 +6,62 @@ Linux GPU/CPU 基础训练推理测试的主程序为`test_train_inference_pytho
 
 - 训练相关：
 
-| 任务类别 | 模型名称 | 单机单卡 | 单机多卡 |
-|  :----: |   :----:  |    :----:  |  :----:   |
-|  变化检测  | BIT | 正常训练 | 正常训练 |
-|  场景分类  | HRNet | 正常训练 | 正常训练 |
-|  目标检测  | PP-YOLO | 正常训练 | 正常训练 |
-|  图像分割  | UNet | 正常训练 | 正常训练 |
+| 任务类别 | 模型名称 | 单机单卡 | 单机多卡 | 参考预测精度 |
+| :----: | :----: | :----: | :----: | :----: |
+|  变化检测  | BIT | 正常训练 | 正常训练 | IoU=71.02% |
+|  变化检测  | CDNet | 正常训练 | 正常训练 | IoU=56.02% |
+|  变化检测  | ChangeFormer | 正常训练 | 正常训练 | IoU=61.65% |
+|  变化检测  | DSAMNet | 正常训练 | 正常训练 | IoU=69.76% |
+|  变化检测  | DSIFN | 正常训练 | 正常训练 | IoU=72.88% |
+|  变化检测  | SNUNet | 正常训练 | 正常训练 | IoU=68.46% |
+|  变化检测  | STANet | 正常训练 | 正常训练 | IoU=65.11% |
+|  变化检测  | FC-EF | 正常训练 | 正常训练 | IoU=64.22% |
+|  变化检测  | FC-Siam-conc | 正常训练 | 正常训练 | IoU=65.79% |
+|  变化检测  | FC-Siam-diff | 正常训练 | 正常训练 | IoU=61.23% |
+|  变化检测  | FCCDN | 正常训练 | 正常训练 | IoU=24.42% |
+|  场景分类  | HRNet | 正常训练 | 正常训练 | Acc(top1)=99.37% |
+|  场景分类  | MobileNetV3 | 正常训练 | 正常训练 | Acc(top1)=99.58% |
+|  场景分类  | ResNet50-vd | 正常训练 | 正常训练 | Acc(top1)=99.26% |
+|  图像复原  | DRN | 正常训练 | 正常训练 | PSNR=24.23 |
+|  图像复原  | ESRGAN | 正常训练 | 正常训练 | PSNR=21.30 |
+|  图像复原  | LESRCNN | 正常训练 | 正常训练 | PSNR=23.18 |
+|  目标检测  | Faster R-CNN | 正常训练 | 正常训练 | mAP=46.99% |
+|  目标检测  | PP-YOLO | 正常训练 | 正常训练 | mAP=56.02% |
+|  目标检测  | PP-YOLO Tiny | 正常训练 | 正常训练 | mAP=44.27% |
+|  目标检测  | PP-YOLOv2 | 正常训练 | 正常训练 | mAP=59.37% |
+|  目标检测  | YOLOv3 | 正常训练 | 正常训练 | mAP=47.33% |
+|  图像分割  | DeepLab V3+ | 正常训练 | 正常训练 | mIoU=56.05% |
+|  图像分割  | UNet | 正常训练 | 正常训练 | mIoU=55.50% |
+
+*注：参考预测精度为whole_train_whole_infer模式下单卡训练汇报的精度数据。*
 
 - 推理相关：
 
 | 任务类别 | 模型名称 | device_CPU | device_GPU | batchsize |
 |  :----:   |  :----: |   :----:   |  :----:  |   :----:   |
-|  变化检测  |  BIT |  支持 | 支持 | 1 |
-|  场景分类  |  HRNet |  支持 | 支持 | 1 |
-|  目标检测  |  YOLO |  支持 | 支持 | 1 |
-|  图像分割  |  UNet |  支持 | 支持 | 1 |
-
+|  变化检测  | BIT | 支持 | 支持 | 1 |
+|  变化检测  | CDNet | 支持 | 支持 | 1 |
+|  变化检测  | ChangeFormer | 支持 | 支持 | 1 |
+|  变化检测  | DSAMNet | 支持 | 支持 | 1 |
+|  变化检测  | DSIFN | 支持 | 支持 | 1 |
+|  变化检测  | SNUNet | 支持 | 支持 | 1 |
+|  变化检测  | STANet | 支持 | 支持 | 1 |
+|  变化检测  | FC-EF | 支持 | 支持 | 1 |
+|  变化检测  | FC-Siam-conc | 支持 | 支持 | 1 |
+|  变化检测  | FC-Siam-diff | 支持 | 支持 | 1 |
+|  场景分类  | HRNet | 支持 | 支持 | 1 |
+|  场景分类  | MobileNetV3 | 支持 | 支持 | 1 |
+|  场景分类  | ResNet50-vd | 支持 | 支持 | 1 |
+|  图像复原  | DRN | 支持 | 支持 | 1 |
+|  图像复原  | ESRGAN | 支持 | 支持 | 1 |
+|  图像复原  | LESRCNN | 支持 | 支持 | 1 |
+|  目标检测  | Faster R-CNN | 支持 | 支持 | 1 |
+|  目标检测  | PP-YOLO | 支持 | 支持 | 1 |
+|  目标检测  | PP-YOLO Tiny | 支持 | 支持 | 1 |
+|  目标检测  | PP-YOLOv2 | 支持 | 支持 | 1 |
+|  目标检测  | YOLOv3 | 支持 | 支持 | 1 |
+|  图像分割  | DeepLab V3+ | 支持 | 支持 | 1 |
+|  图像分割  | UNet | 支持 | 支持 | 1 |
 
 ## 2 测试流程
 
@@ -67,7 +107,7 @@ bash ./test_tipc/test_train_inference_python.sh test_tipc/configs/clas/hrnet/tra
 
 运行相应指令后，在`test_tipc/output`目录中会自动保存运行日志。如lite_train_lite_infer模式下，该目录中可能存在以下文件：
 ```
-test_tipc/output/[task name]/[model name]/
+test_tipc/output/{task name}/{model name}/
 |- results_python.log    # 存储指令执行状态的日志
 |- norm_gpus_0_autocast_null/  # GPU 0号卡上的训练日志和模型保存目录
 ......

test_tipc/infer.py

@@ -13,6 +13,8 @@ from paddle.inference import PrecisionType
 from paddlers.tasks import load_model
 from paddlers.utils import logging
 
+from config_utils import parse_configs
+
 
 class _bool(object):
     def __new__(cls, x):
@@ -101,7 +103,7 @@ class TIPCPredictor(object):
                     logging.warning(
                         "Semantic segmentation models do not support TensorRT acceleration, "
                         "TensorRT is forcibly disabled.")
-                elif 'RCNN' in self._model.__class__.__name__:
+                elif self._model.model_type == 'detector' and 'RCNN' in self._model.__class__.__name__:
                     logging.warning(
                         "RCNN models do not support TensorRT acceleration, "
                         "TensorRT is forcibly disabled.")
@@ -123,7 +125,7 @@ class TIPCPredictor(object):
                     )
                 else:
                     try:
-                        # Cache 10 different shapes for mkldnn to avoid memory leak
+                        # Cache 10 different shapes for mkldnn to avoid memory leak.
                         config.set_mkldnn_cache_capacity(10)
                         config.enable_mkldnn()
                         config.set_cpu_math_library_num_threads(mkl_thread_num)
@@ -158,13 +160,23 @@ class TIPCPredictor(object):
                 'image2': preprocessed_samples[1],
                 'ori_shape': preprocessed_samples[2]
             }
+        elif self._model.model_type == 'restorer':
+            preprocessed_samples = {
+                'image': preprocessed_samples[0],
+                'tar_shape': preprocessed_samples[1]
+            }
         else:
             logging.error(
                 "Invalid model type {}".format(self._model.model_type),
                 exit=True)
         return preprocessed_samples
 
-    def postprocess(self, net_outputs, topk=1, ori_shape=None, transforms=None):
+    def postprocess(self,
+                    net_outputs,
+                    topk=1,
+                    ori_shape=None,
+                    tar_shape=None,
+                    transforms=None):
         if self._model.model_type == 'classifier':
             true_topk = min(self._model.num_classes, topk)
             if self._model.postprocess is None:
@@ -196,6 +208,12 @@ class TIPCPredictor(object):
                 for k, v in zip(['bbox', 'bbox_num', 'mask'], net_outputs)
             }
             preds = self._model.postprocess(net_outputs)
+        elif self._model.model_type == 'restorer':
+            res_maps = self._model.postprocess(
+                net_outputs[0],
+                batch_tar_shape=tar_shape,
+                transforms=transforms.transforms)
+            preds = [{'res_map': res_map} for res_map in res_maps]
         else:
             logging.error(
                 "Invalid model type {}.".format(self._model.model_type),
@@ -232,6 +250,7 @@ class TIPCPredictor(object):
             net_outputs,
             topk,
             ori_shape=preprocessed_input.get('ori_shape', None),
+            tar_shape=preprocessed_input.get('tar_shape', None),
             transforms=transforms)
 
         if self.benchmark and time_it:
@@ -285,7 +304,8 @@ class TIPCPredictor(object):
 if __name__ == '__main__':
     parser = argparse.ArgumentParser()
 
-    parser.add_argument('--file_list', type=str, nargs=2)
+    parser.add_argument('--config', type=str)
+    parser.add_argument('--inherit_off', action='store_true')
     parser.add_argument('--model_dir', type=str, default='./')
     parser.add_argument(
         '--device', type=str, choices=['cpu', 'gpu'], default='cpu')
@@ -300,6 +320,11 @@ if __name__ == '__main__':
 
     args = parser.parse_args()
 
+    cfg = parse_configs(args.config, not args.inherit_off)
+    eval_dataset = cfg['datasets']['eval']
+    data_dir = eval_dataset.args['data_dir']
+    file_list = eval_dataset.args['file_list']
+
     predictor = TIPCPredictor(
         args.model_dir,
         device=args.device,
@@ -310,7 +335,7 @@ if __name__ == '__main__':
         trt_precision_mode=args.precision,
         benchmark=args.benchmark)
 
-    predictor.predict(args.file_list[0], args.file_list[1])
+    predictor.predict(data_dir, file_list)
 
     if args.benchmark:
         predictor.autolog.report()

test_tipc/prepare.sh

@@ -35,6 +35,8 @@ if [[ ${MODE} == 'lite_train_lite_infer' \
         download_and_unzip_dataset "${DATA_DIR}" ucmerced https://paddlers.bj.bcebos.com/datasets/ucmerced.zip
     elif [[ ${task_name} == 'det' ]]; then
         download_and_unzip_dataset "${DATA_DIR}" sarship https://paddlers.bj.bcebos.com/datasets/sarship.zip
+    elif [[ ${task_name} == 'res' ]]; then
+        download_and_unzip_dataset "${DATA_DIR}" rssr https://paddlers.bj.bcebos.com/datasets/rssr_mini.zip
     elif [[ ${task_name} == 'seg' ]]; then
         download_and_unzip_dataset "${DATA_DIR}" rsseg https://paddlers.bj.bcebos.com/datasets/rsseg_mini.zip
     fi
@@ -42,12 +44,26 @@ if [[ ${MODE} == 'lite_train_lite_infer' \
 elif [[ ${MODE} == 'whole_train_whole_infer' ]]; then
 
     if [[ ${task_name} == 'cd' ]]; then
+        rm -rf "${DATA_DIR}/levircd"
         download_and_unzip_dataset "${DATA_DIR}" raw_levircd https://paddlers.bj.bcebos.com/datasets/raw/LEVIR-CD.zip \
         && python tools/prepare_dataset/prepare_levircd.py \
             --in_dataset_dir "${DATA_DIR}/raw_levircd" \
             --out_dataset_dir "${DATA_DIR}/levircd" \
             --crop_size 256 \
             --crop_stride 256
+    elif [[ ${task_name} == 'clas' ]]; then
+        download_and_unzip_dataset "${DATA_DIR}" ucmerced https://paddlers.bj.bcebos.com/datasets/ucmerced.zip
+    elif [[ ${task_name} == 'det' ]]; then
+        rm -rf "${DATA_DIR}/rsod"
+        download_and_unzip_dataset "${DATA_DIR}" raw_rsod https://paddlers.bj.bcebos.com/datasets/raw/RSOD.zip
+        python tools/prepare_dataset/prepare_rsod.py \
+            --in_dataset_dir "${DATA_DIR}/raw_rsod" \
+            --out_dataset_dir "${DATA_DIR}/rsod" \
+            --seed 114514
+    elif [[ ${task_name} == 'res' ]]; then
+        download_and_unzip_dataset "${DATA_DIR}" rssr https://paddlers.bj.bcebos.com/datasets/rssr.zip
+    elif [[ ${task_name} == 'seg' ]]; then
+        download_and_unzip_dataset "${DATA_DIR}" rsseg https://paddlers.bj.bcebos.com/datasets/rsseg.zip
     fi
 
 fi

tests/run_examples.sh

@@ -1 +0,0 @@
-#!/usr/bin/env bash

tests/run_tests.sh

@@ -15,6 +15,3 @@ done
 
 # Test tutorials
 bash run_tutorials.sh
-
-# Test examples
-bash run_examples.sh

tools/prepare_dataset/common.py

@@ -1,4 +1,6 @@
 import argparse
+import random
+import copy
 import os
 import os.path as osp
 from glob import glob
@@ -198,6 +200,20 @@ def create_file_list(file_list, path_tuples, sep=' '):
             f.write(line + '\n')
 
 
+def create_label_list(label_list, labels):
+    """
+    Create label list.
+    
+    Args:
+        label_list (str): Path of label list to create.
+        labels (list[str]|tuple[str]]): Label names.
+    """
+
+    with open(label_list, 'w') as f:
+        for label in labels:
+            f.write(label + '\n')
+
+
 def link_dataset(src, dst):
     """
     Make a symbolic link to a dataset.
@@ -211,5 +227,57 @@ def link_dataset(src, dst):
         raise ValueError(f"{dst} exists and is not a directory.")
     elif not osp.exists(dst):
         os.makedirs(dst)
+    src = osp.realpath(src)
     name = osp.basename(osp.normpath(src))
     os.symlink(src, osp.join(dst, name), target_is_directory=True)
+
+
+def random_split(samples,
+                 ratios=(0.7, 0.2, 0.1),
+                 inplace=True,
+                 drop_remainder=False):
+    """
+    Randomly split the dataset into two or three subsets.
+    
+    Args:
+        samples (list): All samples of the dataset.
+        ratios (tuple[float], optional): If the length of `ratios` is 2,
+            the two elements indicate the ratios of samples used for training
+            and evaluation. If the length of `ratios` is 3, the three elements
+            indicate the ratios of samples used for training, validation, and 
+            testing. Defaults to (0.7, 0.2, 0.1).
+        inplace (bool, optional): Whether to shuffle `samples` in place. 
+            Defaults to True.
+        drop_remainder (bool, optional): Whether to discard the remaining samples.
+            If False, the remaining samples will be included in the last subset.
+            For example, if `ratios` is (0.7, 0.1) and `drop_remainder` is False, 
+            the two subsets after splitting will contain 70% and 30% of the samples, 
+            respectively. Defaults to False.
+    """
+
+    if not inplace:
+        samples = copy.deepcopy(samples)
+
+    if len(samples) == 0:
+        raise ValueError("There are no samples!")
+
+    if len(ratios) not in (2, 3):
+        raise ValueError("`len(ratios)` must be 2 or 3!")
+
+    random.shuffle(samples)
+
+    n_samples = len(samples)
+    acc_r = 0
+    st_idx, ed_idx = 0, 0
+    splits = []
+    for r in ratios:
+        acc_r += r
+        ed_idx = round(acc_r * n_samples)
+        splits.append(samples[st_idx:ed_idx])
+        st_idx = ed_idx
+
+    if ed_idx < len(ratios) and not drop_remainder:
+        # Append remainder to the last split
+        splits[-1].append(splits[ed_idx:])
+
+    return splits

tools/prepare_dataset/prepare_rsod.py

@@ -0,0 +1,56 @@
+#!/usr/bin/env python
+
+import random
+import os.path as osp
+from functools import reduce, partial
+
+from common import (get_default_parser, get_path_tuples, create_file_list,
+                    link_dataset, random_split, create_label_list)
+
+CLASSES = ('aircraft', 'oiltank', 'overpass', 'playground')
+SUBSETS = ('train', 'val', 'test')
+SUBDIRS = ('JPEGImages', osp.sep.join(['Annotation', 'xml']))
+FILE_LIST_PATTERN = "{subset}.txt"
+LABEL_LIST_NAME = "labels.txt"
+URL = ""
+
+if __name__ == '__main__':
+    parser = get_default_parser()
+    parser.add_argument('--seed', type=int, default=None, help="Random seed.")
+    parser.add_argument(
+        '--ratios',
+        type=float,
+        nargs='+',
+        default=(0.7, 0.2, 0.1),
+        help="Ratios of each subset (train/val or train/val/test).")
+    args = parser.parse_args()
+
+    if args.seed is not None:
+        random.seed(args.seed)
+
+    if len(args.ratios) not in (2, 3):
+        raise ValueError("Wrong number of ratios!")
+
+    out_dir = osp.join(args.out_dataset_dir,
+                       osp.basename(osp.normpath(args.in_dataset_dir)))
+
+    link_dataset(args.in_dataset_dir, args.out_dataset_dir)
+
+    splits_list = []
+    for cls in CLASSES:
+        path_tuples = get_path_tuples(
+            *(osp.join(out_dir, cls, subdir) for subdir in SUBDIRS),
+            data_dir=args.out_dataset_dir)
+        splits = random_split(path_tuples, ratios=args.ratios)
+        splits_list.append(splits)
+    splits = map(partial(reduce, list.__add__), zip(*splits_list))
+
+    for subset, split in zip(SUBSETS, splits):
+        file_list = osp.join(
+            args.out_dataset_dir, FILE_LIST_PATTERN.format(subset=subset))
+        create_file_list(file_list, split)
+        print(f"Write file list to {file_list}.")
+
+    label_list = osp.join(args.out_dataset_dir, LABEL_LIST_NAME)
+    create_label_list(label_list, CLASSES)
+    print(f"Write label list to {label_list}.")

tutorials/train/README.md

@@ -12,6 +12,7 @@
 |change_detection/fc_ef.py | 变化检测 | FC-EF |
 |change_detection/fc_siam_conc.py | 变化检测 | FC-Siam-conc |
 |change_detection/fc_siam_diff.py | 变化检测 | FC-Siam-diff |
+|change_detection/fccdn.py | 变化检测 | FCCDN |
 |change_detection/snunet.py | 变化检测 | SNUNet |
 |change_detection/stanet.py | 变化检测 | STANet |
 |classification/hrnet.py | 场景分类 | HRNet |
@@ -22,7 +23,7 @@
 |image_restoration/lesrcnn.py | 图像复原 | LESRCNN |
 |object_detection/faster_rcnn.py | 目标检测 | Faster R-CNN |
 |object_detection/ppyolo.py | 目标检测 | PP-YOLO |
-|object_detection/ppyolotiny.py | 目标检测 | PP-YOLO Tiny |
+|object_detection/ppyolo_tiny.py | 目标检测 | PP-YOLO Tiny |
 |object_detection/ppyolov2.py | 目标检测 | PP-YOLOv2 |
 |object_detection/yolov3.py | 目标检测 | YOLOv3 |
 |semantic_segmentation/deeplabv3p.py | 图像分割 | DeepLab V3+ |

tutorials/train/change_detection/bit.py

@@ -78,7 +78,7 @@ model = pdrs.tasks.cd.BIT()
 
 # 执行模型训练
 model.train(
-    num_epochs=5,
+    num_epochs=10,
     train_dataset=train_dataset,
     train_batch_size=4,
     eval_dataset=eval_dataset,

tutorials/train/change_detection/cdnet.py

@@ -78,7 +78,7 @@ model = pdrs.tasks.cd.CDNet()
 
 # 执行模型训练
 model.train(
-    num_epochs=5,
+    num_epochs=10,
     train_dataset=train_dataset,
     train_batch_size=4,
     eval_dataset=eval_dataset,

tutorials/train/change_detection/changeformer.py

@@ -78,7 +78,7 @@ model = pdrs.tasks.cd.ChangeFormer()
 
 # 执行模型训练
 model.train(
-    num_epochs=5,
+    num_epochs=10,
     train_dataset=train_dataset,
     train_batch_size=4,
     eval_dataset=eval_dataset,

tutorials/train/change_detection/dsamnet.py

@@ -78,7 +78,7 @@ model = pdrs.tasks.cd.DSAMNet()
 
 # 执行模型训练
 model.train(
-    num_epochs=5,
+    num_epochs=10,
     train_dataset=train_dataset,
     train_batch_size=4,
     eval_dataset=eval_dataset,

tutorials/train/change_detection/dsifn.py

@@ -78,7 +78,7 @@ model = pdrs.tasks.cd.DSIFN()
 
 # 执行模型训练
 model.train(
-    num_epochs=5,
+    num_epochs=10,
     train_dataset=train_dataset,
     train_batch_size=4,
     eval_dataset=eval_dataset,

tutorials/train/change_detection/fc_ef.py

@@ -78,7 +78,7 @@ model = pdrs.tasks.cd.FCEarlyFusion()
 
 # 执行模型训练
 model.train(
-    num_epochs=5,
+    num_epochs=10,
     train_dataset=train_dataset,
     train_batch_size=4,
     eval_dataset=eval_dataset,

tutorials/train/change_detection/fc_siam_conc.py

@@ -78,7 +78,7 @@ model = pdrs.tasks.cd.FCSiamConc()
 
 # 执行模型训练
 model.train(
-    num_epochs=5,
+    num_epochs=10,
     train_dataset=train_dataset,
     train_batch_size=4,
     eval_dataset=eval_dataset,

tutorials/train/change_detection/fc_siam_diff.py

@@ -78,7 +78,7 @@ model = pdrs.tasks.cd.FCSiamDiff()
 
 # 执行模型训练
 model.train(
-    num_epochs=5,
+    num_epochs=10,
     train_dataset=train_dataset,
     train_batch_size=4,
     eval_dataset=eval_dataset,

tutorials/train/change_detection/snunet.py

@@ -78,7 +78,7 @@ model = pdrs.tasks.cd.SNUNet()
 
 # 执行模型训练
 model.train(
-    num_epochs=5,
+    num_epochs=10,
     train_dataset=train_dataset,
     train_batch_size=4,
     eval_dataset=eval_dataset,