siton bug

configs/voc/ppyoloe_plus_crn_s_30e_voc.yml

+_BASE_: [
+  '../ppyoloe/ppyoloe_plus_crn_s_80e_coco.yml',
+  '../datasets/voc.yml',
+]
+log_iter: 50
+snapshot_epoch: 5
+weights: output/ppyoloe_plus_crn_s_30e_voc/model_final
+
+pretrain_weights: https://bj.bcebos.com/v1/paddledet/models/ppyoloe_plus_crn_s_80e_coco.pdparams
+depth_mult: 0.33
+width_mult: 0.50
+
+
+TrainReader:
+  batch_size: 8 # default 8 gpus, total bs = 64
+
+EvalReader:
+  batch_size: 4
+
+
+epoch: 30
+LearningRate:
+  base_lr: 0.001
+  schedulers:
+    - !CosineDecay
+      max_epochs: 36
+    - !LinearWarmup
+      start_factor: 0.
+      epochs: 1
+
+
+PPYOLOEHead:
+  static_assigner_epoch: -1
+  nms:
+    name: MultiClassNMS
+    nms_top_k: 1000
+    keep_top_k: 300
+    score_threshold: 0.01
+    nms_threshold: 0.7

configs/voc/yolov5_s_60e_voc.yml

+_BASE_: [
+  '../yolov5/yolov5_s_300e_coco.yml',
+  '../datasets/voc.yml',
+]
+log_iter: 50
+snapshot_epoch: 5
+weights: output/yolov5_s_60e_voc/model_final
+
+pretrain_weights: https://paddledet.bj.bcebos.com/models/yolov5_s_300e_coco.pdparams
+depth_mult: 0.33
+width_mult: 0.50
+
+
+TrainReader:
+  batch_size: 16 # default 8 gpus, total bs = 128
+
+EvalReader:
+  batch_size: 4
+
+
+epoch: 60
+LearningRate:
+  base_lr: 0.001
+  schedulers:
+  - !YOLOv5LRDecay
+    max_epochs: 60
+    min_lr_ratio: 0.01
+  - !ExpWarmup
+    epochs: 1

configs/voc/yolov7_tiny_60e_voc.yml

+_BASE_: [
+  '../yolov7/yolov7_tiny_300e_coco.yml',
+  '../datasets/voc.yml',
+]
+log_iter: 50
+snapshot_epoch: 5
+weights: output/yolov7_tiny_60e_voc/model_final
+
+pretrain_weights: https://paddledet.bj.bcebos.com/models/yolov7_tiny_300e_coco.pdparams
+arch: tiny
+act: LeakyReLU
+
+
+TrainReader:
+  batch_size: 32 # default 8 gpus, total bs = 256
+
+EvalReader:
+  batch_size: 4
+
+
+epoch: 60
+LearningRate:
+  base_lr: 0.001
+  schedulers:
+  - !YOLOv5LRDecay
+    max_epochs: 60
+    min_lr_ratio: 0.1
+  - !ExpWarmup
+    epochs: 1

configs/voc/yolox_s_40e_voc.yml

+_BASE_: [
+  '../yolox/yolox_s_300e_coco.yml',
+  '../datasets/voc.yml',
+]
+log_iter: 50
+snapshot_epoch: 5
+weights: output/yolox_s_40e_voc/model_final
+
+pretrain_weights: https://paddledet.bj.bcebos.com/models/yolox_s_300e_coco.pdparams
+depth_mult: 0.33
+width_mult: 0.50
+
+TrainReader:
+  batch_size: 8 # default 8 gpus, total bs = 64
+
+EvalReader:
+  batch_size: 4
+
+
+epoch: 40
+LearningRate:
+  base_lr: 0.001
+  schedulers:
+  - !CosineDecay
+    max_epochs: 40
+    min_lr_ratio: 0.05
+    last_plateau_epochs: 4
+  - !ExpWarmup
+    epochs: 1

configs/yolof/README.md

+# YOLOF (You Only Look One-level Feature)
+
+## ModelZOO
+
+| 网络网络                | 输入尺寸   | 图片数/GPU | Epochs | 模型推理耗时(ms) | mAP<sup>val<br>0.5:0.95  | Params(M) | FLOPs(G) |    下载链接       | 配置文件 |
+| :--------------------- | :------- | :-------: | :----: | :----------: | :---------------------: | :----------------: |:---------: | :------: |:---------------: |
+| YOLOF-R_50_C5 (paper)  |  800x1333 |    4     |   12    |      -      |         37.7       |  -  |  - | - | - |
+| YOLOF-R_50_C5          |  800x1333 |    4     |   12    |      -      |         38.1       |  44.16  |  241.64 | [下载链接](https://paddledet.bj.bcebos.com/models/yolof_r50_c5_1x_coco.pdparams) | [配置文件](./yolof_r50_c5_1x_coco.yml) |
+
+**注意:**
+  - YOLOF模型训练过程中默认使用8 GPUs进行混合精度训练，总batch_size默认为32。
+
+
+## Citations
+```
+@inproceedings{chen2021you,
+  title={You Only Look One-level Feature},
+  author={Chen, Qiang and Wang, Yingming and Yang, Tong and Zhang, Xiangyu and Cheng, Jian and Sun, Jian},
+  booktitle={IEEE Conference on Computer Vision and Pattern Recognition},
+  year={2021}
+}
+```

configs/yolof/_base_/optimizer_1x.yml

+epoch: 12
+
+LearningRate:
+  base_lr: 0.06
+  schedulers:
+  - !PiecewiseDecay
+    gamma: 0.1
+    milestones: [8, 11]
+  - !LinearWarmup
+    start_factor: 0.00066
+    steps: 1500
+
+OptimizerBuilder:
+  optimizer:
+    momentum: 0.9
+    type: Momentum
+  regularizer:
+    factor: 0.0001
+    type: L2

configs/yolof/_base_/yolof_r50_c5.yml

+architecture: YOLOF
+find_unused_parameters: True
+
+pretrain_weights: https://paddledet.bj.bcebos.com/models/pretrained/ResNet50_cos_pretrained.pdparams
+
+YOLOF:
+  backbone: ResNet
+  neck: DilatedEncoder
+  head: YOLOFHead
+
+ResNet:
+  depth: 50
+  variant: b # resnet-va in paper
+  freeze_at: 0 # res2
+  return_idx: [3] # only res5 feature
+  lr_mult_list: [0.3333, 0.3333, 0.3333, 0.3333]
+
+DilatedEncoder:
+  in_channels: [2048]
+  out_channels: [512]
+  block_mid_channels: 128
+  num_residual_blocks: 4
+  block_dilations: [2, 4, 6, 8]
+
+YOLOFHead:
+  conv_feat:
+    name: YOLOFFeat
+    feat_in: 512
+    feat_out: 512
+    num_cls_convs: 2
+    num_reg_convs: 4
+    norm_type: bn
+  anchor_generator:
+    name: AnchorGenerator
+    anchor_sizes: [[32, 64, 128, 256, 512]]
+    aspect_ratios: [1.0]
+    strides: [32]
+  bbox_assigner:
+    name: UniformAssigner
+    pos_ignore_thr: 0.15
+    neg_ignore_thr: 0.7
+    match_times: 4
+  loss_class:
+    name: FocalLoss
+    gamma: 2.0
+    alpha: 0.25
+  loss_bbox:
+    name: GIoULoss
+  nms:
+    name: MultiClassNMS
+    nms_top_k: 1000
+    keep_top_k: 100
+    score_threshold: 0.05
+    nms_threshold: 0.6

configs/yolof/_base_/yolof_reader.yml

+worker_num: 4
+TrainReader:
+  sample_transforms:
+    - Decode: {}
+    - RandomShift: {prob: 0.5, max_shift: 32}
+    - Resize: {target_size: [800, 1333], keep_ratio: True, interp: 1}
+    - NormalizeImage: {mean: [0.485, 0.456, 0.406], std: [0.229, 0.224, 0.225], is_scale: True}
+    - RandomFlip: {}
+    - Permute: {}
+  batch_transforms:
+    - PadBatch: {pad_to_stride: 32}
+  batch_size: 4
+  shuffle: True
+  drop_last: True
+  collate_batch: False
+
+
+EvalReader:
+  sample_transforms:
+    - Decode: {}
+    - Resize: {target_size: [800, 1333], keep_ratio: True, interp: 1}
+    - NormalizeImage: {is_scale: True, mean: [0.485, 0.456, 0.406], std: [0.229, 0.224, 0.225]}
+    - Permute: {}
+  batch_transforms:
+    - PadBatch: {pad_to_stride: 32}
+  batch_size: 1
+
+
+TestReader:
+  sample_transforms:
+    - Decode: {}
+    - Resize: {target_size: [800, 1333], keep_ratio: True, interp: 1}
+    - NormalizeImage: {is_scale: True, mean: [0.485, 0.456, 0.406], std: [0.229, 0.224, 0.225]}
+    - Permute: {}
+  batch_transforms:
+    - PadBatch: {pad_to_stride: 32}
+  batch_size: 1
+  fuse_normalize: True

configs/yolof/yolof_r50_c5_1x_coco.yml

+_BASE_: [
+  '../datasets/coco_detection.yml',
+  '../runtime.yml',
+  './_base_/optimizer_1x.yml',
+  './_base_/yolof_r50_c5.yml',
+  './_base_/yolof_reader.yml'
+]
+log_iter: 50
+snapshot_epoch: 1
+weights: output/yolof_r50_c5_1x_coco/model_final

configs/yolov3/README.md

+# YOLOv3
+
+## Model Zoo
+
+### YOLOv3 on COCO
+
+| 骨架网络             | 输入尺寸   | 每张GPU图片个数 | 学习率策略 |推理时间(fps) | mAP<sup>val<br>0.5:0.95 |                           下载                          | 配置文件 |
+| :------------------- | :------- | :-----: | :-----: | :------------: | :-----: | :-----------------------------------------------------: | :-----: |
+| DarkNet53(paper)  | 608         |    8    |   270e    |     -     |  33.0  |    -   |    -   |
+| DarkNet53(paper)  | 416         |    8    |   270e    |     -     |  31.0  |    -   |    -   |
+| DarkNet53(paper)  | 320         |    8    |   270e    |     -     |  28.2  |    -   |    -   |
+| DarkNet53         | 608         |    8    |   270e    |     -     |  **39.1**  | [下载链接](https://paddledet.bj.bcebos.com/models/yolov3_darknet53_270e_coco.pdparams) | [配置文件](./yolov3_darknet53_270e_coco.yml) |
+| DarkNet53         | 416         |    8    |   270e    |     -     |  37.7  | [下载链接](https://paddledet.bj.bcebos.com/models/yolov3_darknet53_270e_coco.pdparams) | [配置文件](./yolov3_darknet53_270e_coco.yml) |
+| DarkNet53         | 320         |    8    |   270e    |     -     |  34.8  | [下载链接](https://paddledet.bj.bcebos.com/models/yolov3_darknet53_270e_coco.pdparams) | [配置文件](./yolov3_darknet53_270e_coco.yml) |
+|   ResNet50_vd-DCN    | 608        |    8    |   270e    |     -     |  **40.6**  | [下载链接](https://paddledet.bj.bcebos.com/models/yolov3_r50vd_dcn_270e_coco.pdparams) | [配置文件](./yolov3_r50vd_dcn_270e_coco.yml) |
+|   ResNet50_vd-DCN    | 416        |    8    |   270e    |     -     |  38.2  | [下载链接](https://paddledet.bj.bcebos.com/models/yolov3_r50vd_dcn_270e_coco.pdparams) | [配置文件](./yolov3_r50vd_dcn_270e_coco.yml) |
+|   ResNet50_vd-DCN    | 320        |    8    |   270e    |     -     |  35.1  | [下载链接](https://paddledet.bj.bcebos.com/models/yolov3_r50vd_dcn_270e_coco.pdparams) | [配置文件](./yolov3_r50vd_dcn_270e_coco.yml) |
+| ResNet34         | 608         |    8    |   270e    |     -     |  36.2  | [下载链接](https://paddledet.bj.bcebos.com/models/yolov3_r34_270e_coco.pdparams) | [配置文件](./yolov3_r34_270e_coco.yml) |
+| ResNet34         | 416         |    8    |   270e    |     -     |  34.3  | [下载链接](https://paddledet.bj.bcebos.com/models/yolov3_r34_270e_coco.pdparams) | [配置文件](./yolov3_r34_270e_coco.yml) |
+| ResNet34         | 320         |    8    |   270e    |     -     |  31.2  | [下载链接](https://paddledet.bj.bcebos.com/models/yolov3_r34_270e_coco.pdparams) | [配置文件](./yolov3_r34_270e_coco.yml) |
+| MobileNet-V1         | 608         |    8    |   270e    |     -     |  29.4  | [下载链接](https://paddledet.bj.bcebos.com/models/yolov3_mobilenet_v1_270e_coco.pdparams) | [配置文件](./yolov3_mobilenet_v1_270e_coco.yml) |
+| MobileNet-V1         | 416         |    8    |   270e    |     -     |  29.3  | [下载链接](https://paddledet.bj.bcebos.com/models/yolov3_mobilenet_v1_270e_coco.pdparams) | [配置文件](./yolov3_mobilenet_v1_270e_coco.yml) |
+| MobileNet-V1         | 320         |    8    |   270e    |     -     |  27.2  | [下载链接](https://paddledet.bj.bcebos.com/models/yolov3_mobilenet_v1_270e_coco.pdparams) | [配置文件](./yolov3_mobilenet_v1_270e_coco.yml) |
+| MobileNet-V3         | 608         |    8    |   270e    |     -     |  31.4  | [下载链接](https://paddledet.bj.bcebos.com/models/yolov3_mobilenet_v3_large_270e_coco.pdparams) | [配置文件](./yolov3_mobilenet_v3_large_270e_coco.yml) |
+| MobileNet-V3         | 416         |    8    |   270e    |     -     |  29.6  | [下载链接](https://paddledet.bj.bcebos.com/models/yolov3_mobilenet_v3_large_270e_coco.pdparams) | [配置文件](./yolov3_mobilenet_v3_large_270e_coco.yml) |
+| MobileNet-V3         | 320         |    8    |   270e    |     -     |  27.1  | [下载链接](https://paddledet.bj.bcebos.com/models/yolov3_mobilenet_v3_large_270e_coco.pdparams) | [配置文件](./yolov3_mobilenet_v3_large_270e_coco.yml) |
+| MobileNet-V1-SSLD    | 608         |    8    |   270e    |     -     |  31.0  | [下载链接](https://paddledet.bj.bcebos.com/models/yolov3_mobilenet_v1_ssld_270e_coco.pdparams) | [配置文件](./yolov3_mobilenet_v1_ssld_270e_coco.yml) |
+| MobileNet-V1-SSLD    | 416         |    8    |   270e    |     -     |  30.6  | [下载链接](https://paddledet.bj.bcebos.com/models/yolov3_mobilenet_v1_ssld_270e_coco.pdparams) | [配置文件](./yolov3_mobilenet_v1_ssld_270e_coco.yml) |
+| MobileNet-V1-SSLD    | 320         |    8    |   270e    |     -     |  28.4  | [下载链接](https://paddledet.bj.bcebos.com/models/yolov3_mobilenet_v1_ssld_270e_coco.pdparams) | [配置文件](./yolov3_mobilenet_v1_ssld_270e_coco.yml) |
+
+### YOLOv3 on Pasacl VOC
+
+| 骨架网络     | 输入尺寸 | 每张GPU图片个数 | 学习率策略 |推理时间(fps)| mAP(0.50,11point) | 下载 | 配置文件 |
+| :----------- | :--: | :-----: | :-----: |:------------: |:----: | :-------: | :----: |
+| DarkNet53    | 608  |    8    |   270e  |      -        |  **85.4** (56.1 mAP<br>0.5:0.95)  | [下载链接](https://paddledet.bj.bcebos.com/models/yolov3_darknet53_270e_voc.pdparams) | [配置文件](./yolov3_darknet53_270e_voc.yml) |
+| DarkNet53    | 416  |    8    |   270e  |      -        |  85.2 (57.3 mAP<br>0.5:0.95)  | [下载链接](https://paddledet.bj.bcebos.com/models/yolov3_darknet53_270e_voc.pdparams) | [配置文件](./yolov3_darknet53_270e_voc.yml) |
+| DarkNet53    | 320  |    8    |   270e  |      -        |  84.3 (55.2 mAP<br>0.5:0.95)  | [下载链接](https://paddledet.bj.bcebos.com/models/yolov3_darknet53_270e_voc.pdparams) | [配置文件](./yolov3_darknet53_270e_voc.yml) |
+| MobileNet-V1 | 608  |    8    |   270e  |      -        |  75.2  | [下载链接](https://paddledet.bj.bcebos.com/models/yolov3_mobilenet_v1_270e_voc.pdparams) | [配置文件](./yolov3_mobilenet_v1_270e_voc.yml) |
+| MobileNet-V1 | 416  |    8    |   270e  |      -        |  76.2  | [下载链接](https://paddledet.bj.bcebos.com/models/yolov3_mobilenet_v1_270e_voc.pdparams) | [配置文件](./yolov3_mobilenet_v1_270e_voc.yml) |
+| MobileNet-V1 | 320  |    8    |   270e  |      -        |  74.3  | [下载链接](https://paddledet.bj.bcebos.com/models/yolov3_mobilenet_v1_270e_voc.pdparams) | [配置文件](./yolov3_mobilenet_v1_270e_voc.yml) |
+| MobileNet-V3 | 608  |    8    |   270e  |      -        |  79.6  | [下载链接](https://paddledet.bj.bcebos.com/models/yolov3_mobilenet_v3_large_270e_voc.pdparams) | [配置文件](./yolov3_mobilenet_v3_large_270e_voc.yml) |
+| MobileNet-V3 | 416  |    8    |   270e  |      -        |  78.6  | [下载链接](https://paddledet.bj.bcebos.com/models/yolov3_mobilenet_v3_large_270e_voc.pdparams) | [配置文件](./yolov3_mobilenet_v3_large_270e_voc.yml) |
+| MobileNet-V3 | 320  |    8    |   270e  |      -        |  76.4  | [下载链接](https://paddledet.bj.bcebos.com/models/yolov3_mobilenet_v3_large_270e_voc.pdparams) | [配置文件](./yolov3_mobilenet_v3_large_270e_voc.yml) |
+| MobileNet-V1-SSLD | 608  |    8    |   270e  |      -        |  78.3  | [下载链接](https://paddledet.bj.bcebos.com/models/yolov3_mobilenet_v1_ssld_270e_voc.pdparams) | [配置文件](./yolov3_mobilenet_v1_ssld_270e_voc.yml) |
+| MobileNet-V1-SSLD | 416  |    8    |   270e  |      -        |  79.6  | [下载链接](https://paddledet.bj.bcebos.com/models/yolov3_mobilenet_v1_ssld_270e_voc.pdparams) | [配置文件](./yolov3_mobilenet_v1_ssld_270e_voc.yml) |
+| MobileNet-V1-SSLD | 320  |    8    |   270e  |      -        |  77.3  | [下载链接](https://paddledet.bj.bcebos.com/models/yolov3_mobilenet_v1_ssld_270e_voc.pdparams) | [配置文件](./yolov3_mobilenet_v1_ssld_270e_voc.yml) |
+| MobileNet-V3-SSLD | 608  |    8    |   270e  |      -        |  80.4  | [下载链接](https://paddledet.bj.bcebos.com/models/yolov3_mobilenet_v3_large_ssld_270e_voc.pdparams) | [配置文件](./yolov3_mobilenet_v3_large_ssld_270e_voc.yml) |
+| MobileNet-V3-SSLD | 416  |    8    |   270e  |      -        |  79.2  | [下载链接](https://paddledet.bj.bcebos.com/models/yolov3_mobilenet_v3_large_ssld_270e_voc.pdparams) | [配置文件](./yolov3_mobilenet_v3_large_ssld_270e_voc.yml) |
+| MobileNet-V3-SSLD | 320  |    8    |   270e  |      -        |  77.3  | [下载链接](https://paddledet.bj.bcebos.com/models/yolov3_mobilenet_v3_large_ssld_270e_voc.pdparams) | [配置文件](./yolov3_mobilenet_v3_large_ssld_270e_voc.yml) |
+
+
+**注意:**
+  - YOLOv3模型训练过程中默认使用8 GPUs，总batch_size默认为64，评估时网络尺度默认为`608*608`；
+  - `416*416`和`320*320`尺度只需更改`EvalReader`的`Resize`参数为相应值即可，无需重新训练模型，如：
+  ```
+    EvalReader:
+    sample_transforms:
+        - Decode: {}
+        - Resize: {target_size: [416, 416], keep_ratio: False, interp: 2} # or [320, 320]
+        - NormalizeImage: {mean: [0.485, 0.456, 0.406], std: [0.229, 0.224, 0.225], is_scale: True}
+        - Permute: {}
+    batch_size: 1
+  ```
+  - VOC数据集可以从此[链接](https://bj.bcebos.com/v1/paddledet/data/voc.zip)下载，默认评估指标为mAP(0.50,11point)，如果想转为COCO格式指标的mAP<br>0.5:0.95，可以参照[yolov3_darknet53_270e_voc](./yolov3_darknet53_270e_voc.yml) 添加以下几行重新eval：
+  ```
+    metric: COCO
+    EvalDataset:
+    !COCODataSet
+        image_dir: VOCdevkit/VOC2007/JPEGImages
+        anno_path: voc_test.json
+        dataset_dir: dataset/voc
+  ```
+
+
+## Citations
+```
+@misc{redmon2018yolov3,
+    title={YOLOv3: An Incremental Improvement},
+    author={Joseph Redmon and Ali Farhadi},
+    year={2018},
+    eprint={1804.02767},
+    archivePrefix={arXiv},
+    primaryClass={cs.CV}
+}
+```

configs/yolov3/_base_/optimizer_270e.yml

+epoch: 270
+
+LearningRate:
+  base_lr: 0.001
+  schedulers:
+  - !PiecewiseDecay
+    gamma: 0.1
+    milestones:
+    - 216
+    - 243
+  - !LinearWarmup
+    start_factor: 0.
+    steps: 4000
+
+OptimizerBuilder:
+  optimizer:
+    momentum: 0.9
+    type: Momentum
+  regularizer:
+    factor: 0.0005
+    type: L2

configs/yolov3/_base_/optimizer_40e.yml

+epoch: 40
+
+LearningRate:
+  base_lr: 0.0001
+  schedulers:
+  - name: PiecewiseDecay
+    gamma: 0.1
+    milestones:
+    - 32
+    - 36
+  - name: LinearWarmup
+    start_factor: 0.3333333333333333
+    steps: 100
+
+OptimizerBuilder:
+  optimizer:
+    momentum: 0.9
+    type: Momentum
+  regularizer:
+    factor: 0.0005
+    type: L2

configs/yolov3/_base_/yolov3_darknet53.yml

+architecture: YOLOv3
+pretrain_weights: https://paddledet.bj.bcebos.com/models/pretrained/DarkNet53_pretrained.pdparams
+norm_type: sync_bn
+
+YOLOv3:
+  backbone: DarkNet
+  neck: YOLOv3FPN
+  yolo_head: YOLOv3Head
+  post_process: BBoxPostProcess
+
+DarkNet:
+  depth: 53
+  return_idx: [2, 3, 4]
+
+# use default config
+# YOLOv3FPN:
+
+YOLOv3Head:
+  anchors: [[10, 13], [16, 30], [33, 23],
+            [30, 61], [62, 45], [59, 119],
+            [116, 90], [156, 198], [373, 326]]
+  anchor_masks: [[6, 7, 8], [3, 4, 5], [0, 1, 2]]
+  loss: YOLOv3Loss
+
+YOLOv3Loss:
+  ignore_thresh: 0.7
+  downsample: [32, 16, 8]
+  label_smooth: false
+
+BBoxPostProcess:
+  decode:
+    name: YOLOBox
+    conf_thresh: 0.005
+    downsample_ratio: 32
+    clip_bbox: true
+  nms:
+    name: MultiClassNMS
+    keep_top_k: 100
+    score_threshold: 0.01
+    nms_threshold: 0.45
+    nms_top_k: 1000

configs/yolov3/_base_/yolov3_mobilenet_v1.yml

+architecture: YOLOv3
+pretrain_weights: https://paddledet.bj.bcebos.com/models/pretrained/MobileNetV1_pretrained.pdparams
+norm_type: sync_bn
+
+YOLOv3:
+  backbone: MobileNet
+  neck: YOLOv3FPN
+  yolo_head: YOLOv3Head
+  post_process: BBoxPostProcess
+
+MobileNet:
+  scale: 1
+  feature_maps: [4, 6, 13]
+  with_extra_blocks: false
+  extra_block_filters: []
+
+# use default config
+# YOLOv3FPN:
+
+YOLOv3Head:
+  anchors: [[10, 13], [16, 30], [33, 23],
+            [30, 61], [62, 45], [59, 119],
+            [116, 90], [156, 198], [373, 326]]
+  anchor_masks: [[6, 7, 8], [3, 4, 5], [0, 1, 2]]
+  loss: YOLOv3Loss
+
+YOLOv3Loss:
+  ignore_thresh: 0.7
+  downsample: [32, 16, 8]
+  label_smooth: false
+
+BBoxPostProcess:
+  decode:
+    name: YOLOBox
+    conf_thresh: 0.005
+    downsample_ratio: 32
+    clip_bbox: true
+  nms:
+    name: MultiClassNMS
+    keep_top_k: 100
+    score_threshold: 0.01
+    nms_threshold: 0.45
+    nms_top_k: 1000

configs/yolov3/_base_/yolov3_mobilenet_v3_large.yml

+architecture: YOLOv3
+pretrain_weights: https://paddledet.bj.bcebos.com/models/pretrained/MobileNetV3_large_x1_0_ssld_pretrained.pdparams
+norm_type: sync_bn
+
+YOLOv3:
+  backbone: MobileNetV3
+  neck: YOLOv3FPN
+  yolo_head: YOLOv3Head
+  post_process: BBoxPostProcess
+
+MobileNetV3:
+  model_name: large
+  scale: 1.
+  with_extra_blocks: false
+  extra_block_filters: []
+  feature_maps: [7, 13, 16]
+
+# use default config
+# YOLOv3FPN:
+
+YOLOv3Head:
+  anchors: [[10, 13], [16, 30], [33, 23],
+            [30, 61], [62, 45], [59, 119],
+            [116, 90], [156, 198], [373, 326]]
+  anchor_masks: [[6, 7, 8], [3, 4, 5], [0, 1, 2]]
+  loss: YOLOv3Loss
+
+YOLOv3Loss:
+  ignore_thresh: 0.7
+  downsample: [32, 16, 8]
+  label_smooth: false
+
+BBoxPostProcess:
+  decode:
+    name: YOLOBox
+    conf_thresh: 0.005
+    downsample_ratio: 32
+    clip_bbox: true
+  nms:
+    name: MultiClassNMS
+    keep_top_k: 100
+    score_threshold: 0.01
+    nms_threshold: 0.45
+    nms_top_k: 1000

configs/yolov3/_base_/yolov3_mobilenet_v3_small.yml

+architecture: YOLOv3
+pretrain_weights: https://paddledet.bj.bcebos.com/models/pretrained/MobileNetV3_small_x1_0_ssld_pretrained.pdparams
+norm_type: sync_bn
+
+YOLOv3:
+  backbone: MobileNetV3
+  neck: YOLOv3FPN
+  yolo_head: YOLOv3Head
+  post_process: BBoxPostProcess
+
+MobileNetV3:
+  model_name: small
+  scale: 1.
+  with_extra_blocks: false
+  extra_block_filters: []
+  feature_maps: [4, 9, 12]
+
+# use default config
+# YOLOv3FPN:
+
+YOLOv3Head:
+  anchors: [[10, 13], [16, 30], [33, 23],
+            [30, 61], [62, 45], [59, 119],
+            [116, 90], [156, 198], [373, 326]]
+  anchor_masks: [[6, 7, 8], [3, 4, 5], [0, 1, 2]]
+  loss: YOLOv3Loss
+
+YOLOv3Loss:
+  ignore_thresh: 0.7
+  downsample: [32, 16, 8]
+  label_smooth: false
+
+BBoxPostProcess:
+  decode:
+    name: YOLOBox
+    conf_thresh: 0.005
+    downsample_ratio: 32
+    clip_bbox: true
+  nms:
+    name: MultiClassNMS
+    keep_top_k: 100
+    score_threshold: 0.01
+    nms_threshold: 0.45
+    nms_top_k: 1000

configs/yolov3/_base_/yolov3_r34.yml

+architecture: YOLOv3
+pretrain_weights: https://paddledet.bj.bcebos.com/models/pretrained/ResNet34_pretrained.pdparams
+norm_type: sync_bn
+
+YOLOv3:
+  backbone: ResNet
+  neck: YOLOv3FPN
+  yolo_head: YOLOv3Head
+  post_process: BBoxPostProcess
+
+ResNet:
+  depth: 34
+  return_idx: [1, 2, 3]
+  freeze_at: -1
+  freeze_norm: false
+  norm_decay: 0.
+
+YOLOv3Head:
+  anchors: [[10, 13], [16, 30], [33, 23],
+            [30, 61], [62, 45], [59, 119],
+            [116, 90], [156, 198], [373, 326]]
+  anchor_masks: [[6, 7, 8], [3, 4, 5], [0, 1, 2]]
+  loss: YOLOv3Loss
+
+YOLOv3Loss:
+  ignore_thresh: 0.7
+  downsample: [32, 16, 8]
+  label_smooth: false
+
+BBoxPostProcess:
+  decode:
+    name: YOLOBox
+    conf_thresh: 0.005
+    downsample_ratio: 32
+    clip_bbox: true
+  nms:
+    name: MultiClassNMS
+    keep_top_k: 100
+    score_threshold: 0.01
+    nms_threshold: 0.45
+    nms_top_k: 1000

configs/yolov3/_base_/yolov3_r50vd_dcn.yml

+architecture: YOLOv3
+pretrain_weights: https://paddledet.bj.bcebos.com/models/pretrained/ResNet50_vd_ssld_pretrained.pdparams
+norm_type: sync_bn
+
+YOLOv3:
+  backbone: ResNet
+  neck: YOLOv3FPN
+  yolo_head: YOLOv3Head
+  post_process: BBoxPostProcess
+
+ResNet:
+  depth: 50
+  variant: d
+  return_idx: [1, 2, 3]
+  dcn_v2_stages: [3]
+  freeze_at: -1
+  freeze_norm: false
+  norm_decay: 0.
+
+# YOLOv3FPN:
+
+YOLOv3Head:
+  anchors: [[10, 13], [16, 30], [33, 23],
+            [30, 61], [62, 45], [59, 119],
+            [116, 90], [156, 198], [373, 326]]
+  anchor_masks: [[6, 7, 8], [3, 4, 5], [0, 1, 2]]
+  loss: YOLOv3Loss
+
+YOLOv3Loss:
+  ignore_thresh: 0.7
+  downsample: [32, 16, 8]
+  label_smooth: false
+
+BBoxPostProcess:
+  decode:
+    name: YOLOBox
+    conf_thresh: 0.005
+    downsample_ratio: 32
+    clip_bbox: true
+  nms:
+    name: MultiClassNMS
+    keep_top_k: 100
+    score_threshold: 0.01
+    nms_threshold: 0.45
+    nms_top_k: 1000

configs/yolov3/_base_/yolov3_reader.yml

+worker_num: 2
+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]}
+  batch_size: 8
+  shuffle: true
+  drop_last: true
+  mixup_epoch: 250
+  use_shared_memory: true
+
+EvalReader:
+  inputs_def:
+    num_max_boxes: 50
+  sample_transforms:
+    - Decode: {}
+    - Resize: {target_size: [608, 608], keep_ratio: False, interp: 2}
+    - NormalizeImage: {mean: [0.485, 0.456, 0.406], std: [0.229, 0.224, 0.225], is_scale: True}
+    - Permute: {}
+  batch_size: 1
+
+TestReader:
+  inputs_def:
+    image_shape: [3, 608, 608]
+  sample_transforms:
+    - Decode: {}
+    - Resize: {target_size: [608, 608], keep_ratio: False, interp: 2}
+    - NormalizeImage: {mean: [0.485, 0.456, 0.406], std: [0.229, 0.224, 0.225], is_scale: True}
+    - Permute: {}
+  batch_size: 1

configs/yolov3/yolov3_darknet53_270e_coco.yml

+_BASE_: [
+  '../datasets/coco_detection.yml',
+  '../runtime.yml',
+  '_base_/optimizer_270e.yml',
+  '_base_/yolov3_darknet53.yml',
+  '_base_/yolov3_reader.yml',
+]
+
+snapshot_epoch: 5
+weights: output/yolov3_darknet53_270e_coco/model_final