[cherry-pick]refine yolov3 configs (#2758)

* refine yolov3 configs

* fix dataset problem

configs/datasets/roadsign_voc.yml

 metric: VOC
-map_type: 11point
+map_type: integral
 num_classes: 4
 
 TrainDataset:

configs/yolov3/yolov3_mobilenet_v1_270e_voc.yml

@@ -9,41 +9,6 @@ _BASE_: [
 snapshot_epoch: 5
 weights: output/yolov3_mobilenet_v1_270e_voc/model_final
 
-TrainReader:
-  inputs_def:
-    num_max_boxes: 50
-  sample_transforms:
-    - Decode: {}
-    - Mixup: {alpha: 1.5, beta: 1.5}
-    - RandomDistort: {}
-    - RandomExpand: {fill_value: [123.675, 116.28, 103.53]}
-    - RandomCrop: {}
-    - RandomFlip: {}
-  batch_transforms:
-    - BatchRandomResize:
-        target_size: [320, 352, 384, 416, 448, 480, 512, 544, 576, 608]
-        random_size: True
-        random_interp: True
-        keep_ratio: False
-    - NormalizeBox: {}
-    - PadBox: {num_max_boxes: 50}
-    - BboxXYXY2XYWH: {}
-    - NormalizeImage: {mean: [0.485, 0.456, 0.406], std: [0.229, 0.224, 0.225], is_scale: True}
-    - Permute: {}
-    - Gt2YoloTarget:
-        anchor_masks: [[6, 7, 8], [3, 4, 5], [0, 1, 2]]
-        anchors: [[10, 13], [16, 30], [33, 23], [30, 61], [62, 45], [59, 119], [116, 90], [156, 198], [373, 326]]
-        downsample_ratios: [32, 16, 8]
-        num_classes: 20
-  batch_size: 8
-  shuffle: true
-  drop_last: true
-  mixup_epoch: 250
-
-EvalReader:
-  batch_transforms:
-  - PadBatch: {pad_gt: True}
-
 LearningRate:
   base_lr: 0.001
   schedulers:

configs/yolov3/yolov3_mobilenet_v1_roadsign.yml

@@ -5,45 +5,12 @@ _BASE_: [
   '_base_/yolov3_reader.yml',
 ]
 pretrain_weights: https://paddledet.bj.bcebos.com/models/yolov3_mobilenet_v1_270e_coco.pdparams
-norm_type: sync_bn
 weights: output/yolov3_mobilenet_v1_roadsign/model_final
-metric: VOC
-map_type: integral
 
 YOLOv3Loss:
   ignore_thresh: 0.7
   label_smooth: true
 
-TrainReader:
-  inputs_def:
-    num_max_boxes: 50
-  sample_transforms:
-    - Decode: {}
-    - Mixup: {alpha: 1.5, beta: 1.5}
-    - RandomDistort: {}
-    - RandomExpand: {fill_value: [123.675, 116.28, 103.53]}
-    - RandomCrop: {}
-    - RandomFlip: {}
-  batch_transforms:
-    - BatchRandomResize:
-        target_size: [320, 352, 384, 416, 448, 480, 512, 544, 576, 608]
-        random_size: True
-        random_interp: True
-        keep_ratio: False
-    - NormalizeBox: {}
-    - PadBox: {num_max_boxes: 50}
-    - BboxXYXY2XYWH: {}
-    - NormalizeImage: {mean: [0.485, 0.456, 0.406], std: [0.229, 0.224, 0.225], is_scale: True}
-    - Permute: {}
-    - Gt2YoloTarget:
-        anchor_masks: [[6, 7, 8], [3, 4, 5], [0, 1, 2]]
-        anchors: [[10, 13], [16, 30], [33, 23], [30, 61], [62, 45], [59, 119], [116, 90], [156, 198], [373, 326]]
-        downsample_ratios: [32, 16, 8]
-        num_classes: 4
-  batch_size: 8
-  shuffle: true
-  drop_last: true
-
 snapshot_epoch: 2
 epoch: 40
 

configs/yolov3/yolov3_mobilenet_v1_ssld_270e_voc.yml

 _BASE_: [
-  '../datasets/coco_detection.yml',
+  '../datasets/voc.yml',
   '../runtime.yml',
   '_base_/optimizer_270e.yml',
   '_base_/yolov3_mobilenet_v1.yml',
@@ -10,41 +10,6 @@ snapshot_epoch: 5
 pretrain_weights:  https://paddledet.bj.bcebos.com/models/pretrained/MobileNetV1_ssld_pretrained.pdparams
 weights: output/yolov3_mobilenet_v1_ssld_270e_voc/model_final
 
-TrainReader:
-  inputs_def:
-    num_max_boxes: 50
-  sample_transforms:
-    - Decode: {}
-    - Mixup: {alpha: 1.5, beta: 1.5}
-    - RandomDistort: {}
-    - RandomExpand: {fill_value: [123.675, 116.28, 103.53]}
-    - RandomCrop: {}
-    - RandomFlip: {}
-  batch_transforms:
-    - BatchRandomResize:
-        target_size: [320, 352, 384, 416, 448, 480, 512, 544, 576, 608]
-        random_size: True
-        random_interp: True
-        keep_ratio: False
-    - NormalizeBox: {}
-    - PadBox: {num_max_boxes: 50}
-    - BboxXYXY2XYWH: {}
-    - NormalizeImage: {mean: [0.485, 0.456, 0.406], std: [0.229, 0.224, 0.225], is_scale: True}
-    - Permute: {}
-    - Gt2YoloTarget:
-        anchor_masks: [[6, 7, 8], [3, 4, 5], [0, 1, 2]]
-        anchors: [[10, 13], [16, 30], [33, 23], [30, 61], [62, 45], [59, 119], [116, 90], [156, 198], [373, 326]]
-        downsample_ratios: [32, 16, 8]
-        num_classes: 20
-  batch_size: 8
-  shuffle: true
-  drop_last: true
-  mixup_epoch: 250
-
-EvalReader:
-  batch_transforms:
-  - PadBatch: {pad_gt: True}
-
 LearningRate:
   base_lr: 0.001
   schedulers:

configs/yolov3/yolov3_mobilenet_v3_large_270e_voc.yml

@@ -9,41 +9,6 @@ _BASE_: [
 snapshot_epoch: 5
 weights: output/yolov3_mobilenet_v3_large_270e_voc/model_final
 
-TrainReader:
-  inputs_def:
-    num_max_boxes: 50
-  sample_transforms:
-    - Decode: {}
-    - Mixup: {alpha: 1.5, beta: 1.5}
-    - RandomDistort: {}
-    - RandomExpand: {fill_value: [123.675, 116.28, 103.53]}
-    - RandomCrop: {}
-    - RandomFlip: {}
-  batch_transforms:
-    - BatchRandomResize:
-        target_size: [320, 352, 384, 416, 448, 480, 512, 544, 576, 608]
-        random_size: True
-        random_interp: True
-        keep_ratio: False
-    - NormalizeBox: {}
-    - PadBox: {num_max_boxes: 50}
-    - BboxXYXY2XYWH: {}
-    - NormalizeImage: {mean: [0.485, 0.456, 0.406], std: [0.229, 0.224, 0.225], is_scale: True}
-    - Permute: {}
-    - Gt2YoloTarget:
-        anchor_masks: [[6, 7, 8], [3, 4, 5], [0, 1, 2]]
-        anchors: [[10, 13], [16, 30], [33, 23], [30, 61], [62, 45], [59, 119], [116, 90], [156, 198], [373, 326]]
-        downsample_ratios: [32, 16, 8]
-        num_classes: 20
-  batch_size: 8
-  shuffle: true
-  drop_last: true
-  mixup_epoch: 250
-
-EvalReader:
-  batch_transforms:
-  - PadBatch: {pad_gt: True}
-
 LearningRate:
   base_lr: 0.001
   schedulers:

configs/yolov3/yolov3_mobilenet_v3_large_ssld_270e_voc.yml

 _BASE_: [
-  '../datasets/coco_detection.yml',
+  '../datasets/voc.yml',
   '../runtime.yml',
   '_base_/optimizer_270e.yml',
   '_base_/yolov3_mobilenet_v3_large.yml',
@@ -10,41 +10,6 @@ snapshot_epoch: 5
 pretrain_weights:  https://paddledet.bj.bcebos.com/models/pretrained/MobileNetV3_large_x1_0_ssld_pretrained.pdparams
 weights: output/yolov3_mobilenet_v3_large_ssld_270e_voc/model_final
 
-TrainReader:
-  inputs_def:
-    num_max_boxes: 50
-  sample_transforms:
-    - Decode: {}
-    - Mixup: {alpha: 1.5, beta: 1.5}
-    - RandomDistort: {}
-    - RandomExpand: {fill_value: [123.675, 116.28, 103.53]}
-    - RandomCrop: {}
-    - RandomFlip: {}
-  batch_transforms:
-    - BatchRandomResize:
-        target_size: [320, 352, 384, 416, 448, 480, 512, 544, 576, 608]
-        random_size: True
-        random_interp: True
-        keep_ratio: False
-    - NormalizeBox: {}
-    - PadBox: {num_max_boxes: 50}
-    - BboxXYXY2XYWH: {}
-    - NormalizeImage: {mean: [0.485, 0.456, 0.406], std: [0.229, 0.224, 0.225], is_scale: True}
-    - Permute: {}
-    - Gt2YoloTarget:
-        anchor_masks: [[6, 7, 8], [3, 4, 5], [0, 1, 2]]
-        anchors: [[10, 13], [16, 30], [33, 23], [30, 61], [62, 45], [59, 119], [116, 90], [156, 198], [373, 326]]
-        downsample_ratios: [32, 16, 8]
-        num_classes: 20
-  batch_size: 8
-  shuffle: true
-  drop_last: true
-  mixup_epoch: 250
-
-EvalReader:
-  batch_transforms:
-  - PadBatch: {pad_gt: True}
-
 LearningRate:
   base_lr: 0.001
   schedulers: