Add models basd on OpenImage-v5 and Object365 (#26)

* Add cascade_rcnn_cls_aware_r200_vd_fpn_dcnv2_nonlocal_softnms model on oidv5 and obj365.
* Add obj365 config annotation
* Update docs.

README.md

@@ -71,6 +71,7 @@ PaddleDetection的目的是为工业界和学术界提供丰富、易用的目
 - [行人检测和车辆检测预训练模型](contrib/README_cn.md) 针对不同场景的检测模型
 - [YOLOv3增强模型](docs/YOLOv3_ENHANCEMENT.md) 改进原始YOLOv3，精度达到41.4%，原论文精度为33.0%，同时预测速度也得到提升
 - [Objects365 2019 Challenge夺冠模型](docs/CACascadeRCNN.md) Objects365 Full Track任务中最好的单模型之一,精度达到31.7%
+- [Open Images V5和Objects365数据集模型](docs/OIDV5_BASELINE_MODEL.md)
 
 
 ## 模型压缩
@@ -90,8 +91,13 @@ PaddleDetection的目的是为工业界和学术界提供丰富、易用的目
 
 ## 版本更新
 
-### 10/2019
+### 21/11/2019
+- 增加CascadeClsAware RCNN模型。
+- 增加CBNet，ResNet200和Non-local模型。
+- 增加SoftNMS。
+- 增加Open Image V5数据集和Objects365数据集模型。
 
+### 10/2019
 - 增加增强版YOLOv3模型，精度高达41.4%。
 - 增加人脸检测模型BlazeFace、Faceboxes。
 - 丰富基于COCO的模型，精度高达51.9%。

README_en.md

@@ -80,6 +80,7 @@ Advanced Features:
 - [Pretrained models for pedestrian  and vehicle detection](contrib/README.md) Models for object detection in specific scenarios. 
 - [YOLOv3 enhanced model](docs/YOLOv3_ENHANCEMENT.md) Compared to MAP of 33.0% in paper, enhanced YOLOv3 reaches the MAP of 41.4% and inference speed is improved as well
 - [Objects365 2019 Challenge champion model](docs/CACascadeRCNN.md) One of the best single models in Objects365 Full Track of which MAP reaches 31.7%.
+- [Open Images Dataset V5 and Objects365 Dataset models](docs/OIDV5_BASELINE_MODEL.md)
 
 ## Model compression
 
@@ -98,6 +99,12 @@ Advanced Features:
 
 ## Updates
 
+#### 21/11/2019
+- Add CascadeClsAware RCNN model.
+- Add CBNet, ResNet200 and Non-local model.
+- Add SoftNMS.
+- Add models of Open Images Dataset V5 and Objects365 Dataset.
+
 #### 10/2019
 
 - Add enhanced YOLOv3 models, box mAP up to 41.4%.

configs/obj365/cascade_rcnn_cls_aware_r200_vd_fpn_dcnv2_nonlocal_softnms.yml

+architecture: CascadeRCNNClsAware
+train_feed: FasterRCNNTrainFeed
+eval_feed: FasterRCNNEvalFeed
+test_feed: FasterRCNNTestFeed
+max_iters: 800000
+snapshot_iter: 10000
+use_gpu: true
+log_smooth_window: 20
+save_dir: output
+pretrain_weights: https://paddle-imagenet-models-name.bj.bcebos.com/ResNet200_vd_pretrained.tar
+weights: output/cascade_rcnn_cls_aware_r200_vd_fpn_dcnv2_nonlocal_softnms/model_final
+# obj365 dataset format and its eval method are same as those for coco
+metric: COCO
+num_classes: 366
+
+CascadeRCNNClsAware:
+  backbone: ResNet
+  fpn: FPN
+  rpn_head: FPNRPNHead
+  roi_extractor: FPNRoIAlign
+  bbox_head: CascadeBBoxHead
+  bbox_assigner: CascadeBBoxAssigner
+
+ResNet:
+  norm_type: bn
+  depth: 200
+  feature_maps: [2, 3, 4, 5]
+  freeze_at: 2
+  variant: d
+  dcn_v2_stages: [3, 4, 5]
+  nonlocal_stages: [4]
+
+FPN:
+  min_level: 2
+  max_level: 6
+  num_chan: 256
+  spatial_scale: [0.03125, 0.0625, 0.125, 0.25]
+
+FPNRPNHead:
+  anchor_generator:
+    anchor_sizes: [32, 64, 128, 256, 512]
+    aspect_ratios: [0.5, 1.0, 2.0]
+    stride: [16.0, 16.0]
+    variance: [1.0, 1.0, 1.0, 1.0]
+  anchor_start_size: 32
+  min_level: 2
+  max_level: 6
+  num_chan: 256
+  rpn_target_assign:
+    rpn_batch_size_per_im: 256
+    rpn_fg_fraction: 0.5
+    rpn_positive_overlap: 0.7
+    rpn_negative_overlap: 0.3
+    rpn_straddle_thresh: 0.0
+  train_proposal:
+    min_size: 0.0
+    nms_thresh: 0.7
+    pre_nms_top_n: 2000
+    post_nms_top_n: 2000
+  test_proposal:
+    min_size: 0.0
+    nms_thresh: 0.7
+    pre_nms_top_n: 1000
+    post_nms_top_n: 1000
+
+FPNRoIAlign:
+  canconical_level: 4
+  canonical_size: 224
+  min_level: 2
+  max_level: 5
+  box_resolution: 14
+  sampling_ratio: 2
+
+CascadeBBoxAssigner:
+  batch_size_per_im: 512
+  bbox_reg_weights: [10, 20, 30]
+  bg_thresh_lo: [0.0, 0.0, 0.0]
+  bg_thresh_hi: [0.5, 0.6, 0.7]
+  fg_thresh: [0.5, 0.6, 0.7]
+  fg_fraction: 0.25
+  class_aware: True
+
+CascadeBBoxHead:
+  head: CascadeTwoFCHead
+  nms: MultiClassSoftNMS
+
+CascadeTwoFCHead:
+  mlp_dim: 1024
+
+MultiClassSoftNMS:
+  score_threshold: 0.001
+  keep_top_k: 300
+  softnms_sigma: 0.15
+
+LearningRate:
+  base_lr: 0.01
+  schedulers:
+  - !PiecewiseDecay
+    gamma: 0.1
+    milestones: [520000, 740000]
+  - !LinearWarmup
+    start_factor: 0.1
+    steps: 1000
+
+OptimizerBuilder:
+  optimizer:
+    momentum: 0.9
+    type: Momentum
+  regularizer:
+    factor: 0.0001
+    type: L2
+
+FasterRCNNTrainFeed:
+  batch_size: 1
+  dataset:
+    dataset_dir: dataset/obj365
+    annotation: train.json
+    image_dir: train
+  sample_transforms: 
+  - !DecodeImage
+    to_rgb: True
+    with_mixup: False
+  - !RandomFlipImage
+    prob: 0.5
+  - !NormalizeImage
+    is_channel_first: false
+    is_scale: True
+    mean:
+    - 0.485
+    - 0.456
+    - 0.406
+    std:
+    - 0.229
+    - 0.224
+    - 0.225 
+  - !ResizeImage
+    interp: 1
+    target_size: [416, 448, 480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800, 832, 864, 896, 928, 960, 992, 1024, 1056, 1088, 1120, 1152, 1184, 1216, 1248, 1280, 1312, 1344, 1376, 1408]
+    max_size: 1800
+    use_cv2: true
+  - !Permute
+    to_bgr: false
+  batch_transforms:
+  - !PadBatch
+    pad_to_stride: 32
+  drop_last: false
+  num_workers: 2
+
+FasterRCNNEvalFeed:
+  batch_size: 1
+  dataset:
+    dataset_dir: dataset/obj365
+    annotation: val.json
+    image_dir: val
+  sample_transforms: 
+  - !DecodeImage
+    to_rgb: True
+    with_mixup: False
+  - !NormalizeImage
+    is_channel_first: false
+    is_scale: True
+    mean:
+    - 0.485
+    - 0.456
+    - 0.406
+    std:
+    - 0.229
+    - 0.224
+    - 0.225 
+  - !ResizeImage
+    interp: 1
+    target_size:
+    - 1200
+    max_size: 2000
+    use_cv2: true
+  - !Permute
+    to_bgr: false
+  batch_transforms:
+  - !PadBatch
+    pad_to_stride: 32
+
+FasterRCNNTestFeed:
+  batch_size: 1
+  dataset:
+    annotation: dataset/obj365/val.json
+  batch_transforms:
+  - !PadBatch
+    pad_to_stride: 32
+  drop_last: false
+  num_workers: 2

configs/oidv5/cascade_rcnn_cls_aware_r200_vd_fpn_dcnv2_nonlocal_softnms.yml

+architecture: CascadeRCNNClsAware
+train_feed: FasterRCNNTrainFeed
+eval_feed: FasterRCNNEvalFeed
+test_feed: FasterRCNNTestFeed
+max_iters: 1500000
+snapshot_iter: 10000
+use_gpu: true
+log_smooth_window: 20
+save_dir: output
+pretrain_weights: https://paddle-imagenet-models-name.bj.bcebos.com/ResNet200_vd_pretrained.tar
+weights: output/cascade_rcnn_cls_aware_r200_vd_fpn_dcnv2_nonlocal_softnms/model_final
+metric: OID
+num_classes: 501
+
+CascadeRCNNClsAware:
+  backbone: ResNet
+  fpn: FPN
+  rpn_head: FPNRPNHead
+  roi_extractor: FPNRoIAlign
+  bbox_head: CascadeBBoxHead
+  bbox_assigner: CascadeBBoxAssigner
+
+ResNet:
+  norm_type: bn
+  depth: 200
+  feature_maps: [2, 3, 4, 5]
+  freeze_at: 2
+  variant: d
+  dcn_v2_stages: [3, 4, 5]
+  nonlocal_stages: [4]
+
+FPN:
+  min_level: 2
+  max_level: 6
+  num_chan: 256
+  spatial_scale: [0.03125, 0.0625, 0.125, 0.25]
+
+FPNRPNHead:
+  anchor_generator:
+    anchor_sizes: [32, 64, 128, 256, 512]
+    aspect_ratios: [0.5, 1.0, 2.0]
+    stride: [16.0, 16.0]
+    variance: [1.0, 1.0, 1.0, 1.0]
+  anchor_start_size: 32
+  min_level: 2
+  max_level: 6
+  num_chan: 256
+  rpn_target_assign:
+    rpn_batch_size_per_im: 256
+    rpn_fg_fraction: 0.5
+    rpn_positive_overlap: 0.7
+    rpn_negative_overlap: 0.3
+    rpn_straddle_thresh: 0.0
+  train_proposal:
+    min_size: 0.0
+    nms_thresh: 0.7
+    pre_nms_top_n: 2000
+    post_nms_top_n: 2000
+  test_proposal:
+    min_size: 0.0
+    nms_thresh: 0.7
+    pre_nms_top_n: 1000
+    post_nms_top_n: 1000
+
+FPNRoIAlign:
+  canconical_level: 4
+  canonical_size: 224
+  min_level: 2
+  max_level: 5
+  box_resolution: 14
+  sampling_ratio: 2
+
+CascadeBBoxAssigner:
+  batch_size_per_im: 512
+  bbox_reg_weights: [10, 20, 30]
+  bg_thresh_lo: [0.0, 0.0, 0.0]
+  bg_thresh_hi: [0.5, 0.6, 0.7]
+  fg_thresh: [0.5, 0.6, 0.7]
+  fg_fraction: 0.25
+  class_aware: True
+
+CascadeBBoxHead:
+  head: CascadeTwoFCHead
+  nms: MultiClassSoftNMS
+
+CascadeTwoFCHead:
+  mlp_dim: 1024
+
+MultiClassSoftNMS:
+  score_threshold: 0.001
+  keep_top_k: 300
+  softnms_sigma: 0.15
+
+LearningRate:
+  base_lr: 0.01
+  schedulers:
+  - !PiecewiseDecay
+    gamma: 0.1
+    milestones: [1000000, 1400000]
+  - !LinearWarmup
+    start_factor: 0.1
+    steps: 1000
+
+OptimizerBuilder:
+  optimizer:
+    momentum: 0.9
+    type: Momentum
+  regularizer:
+    factor: 0.0001
+    type: L2
+
+FasterRCNNTrainFeed:
+  batch_size: 1
+  dataset:
+    dataset_dir: dataset/oid
+    annotation: train.json
+    image_dir: train
+  sample_transforms: 
+  - !DecodeImage
+    to_rgb: True
+    with_mixup: False
+  - !RandomFlipImage
+    prob: 0.5
+  - !NormalizeImage
+    is_channel_first: false
+    is_scale: True
+    mean:
+    - 0.485
+    - 0.456
+    - 0.406
+    std:
+    - 0.229
+    - 0.224
+    - 0.225 
+  - !ResizeImage
+    interp: 1
+    target_size: [416, 448, 480, 512, 544, 576, 608, 640, 672, 704, 736, 768, 800, 832, 864, 896, 928, 960, 992, 1024, 1056, 1088, 1120, 1152, 1184, 1216, 1248, 1280, 1312, 1344, 1376, 1408]
+    max_size: 1800
+    use_cv2: true
+  - !Permute
+    to_bgr: false
+  batch_transforms:
+  - !PadBatch
+    pad_to_stride: 32
+  drop_last: false
+  num_workers: 2
+
+FasterRCNNEvalFeed:
+  batch_size: 1
+  dataset:
+    dataset_dir: dataset/oidv5
+    annotation: val.json
+    image_dir: val
+  sample_transforms: 
+  - !DecodeImage
+    to_rgb: True
+    with_mixup: False
+  - !NormalizeImage
+    is_channel_first: false
+    is_scale: True
+    mean:
+    - 0.485
+    - 0.456
+    - 0.406
+    std:
+    - 0.229
+    - 0.224
+    - 0.225 
+  - !ResizeImage
+    interp: 1
+    target_size:
+    - 1200
+    max_size: 2000
+    use_cv2: true
+  - !Permute
+    to_bgr: false
+  batch_transforms:
+  - !PadBatch
+    pad_to_stride: 32
+
+FasterRCNNTestFeed:
+  batch_size: 1
+  dataset:
+    annotation: dataset/oidv5/val.json
+  batch_transforms:
+  - !PadBatch
+    pad_to_stride: 32
+  drop_last: false
+  num_workers: 2

docs/MODEL_ZOO.md

@@ -177,3 +177,8 @@ randomly cropping, randomly expansion, randomly flipping.
 ## Face Detection
 
 Please refer [face detection models](../configs/face_detection) for details.
+
+
+## Object Detection in Open Images Dataset V5
+
+Please refer [Open Images Dataset V5 Baseline model](OIDV5_BASELINE_MODEL.md) for details.

docs/MODEL_ZOO_cn.md

@@ -83,6 +83,7 @@ Paddle提供基于ImageNet的骨架网络预训练模型。所有预训练模型
 | CBResNet200-vd-FPN-Nonlocal   | Cascade Faster  | c3-c5 |     1     |   2.5x    |     -     |  53.3%(softnms)  |    -    | [下载链接](https://paddlemodels.bj.bcebos.com/object_detection/cascade_rcnn_cbr200_vd_fpn_dcnv2_nonlocal_softnms.tar) |
 
 
+
 #### 注意事项:
 - Deformable卷积网络v2(dcn_v2)参考自论文[Deformable ConvNets v2](https://arxiv.org/abs/1811.11168).
 - `c3-c5`意思是在resnet模块的3到5阶段增加`dcn`.
@@ -166,4 +167,9 @@ Paddle提供基于ImageNet的骨架网络预训练模型。所有预训练模型
 
 ## 人脸检测
 
-详细请参考[人脸检测模型](../configs/face_detection).
+详细请参考[人脸检测模型](../configs/face_detection)。
+
+
+## 基于Open Images V5数据集的物体检测
+
+详细请参考[Open Images V5数据集基线模型](OIDV5_BASELINE_MODEL.md)。

docs/OIDV5_BASELINE_MODEL.md

+# CascadeCA RCNN
+## 简介
+CascadeCA RCNN是百度视觉技术部在Google AI Open Images 2019-Object Detction比赛中的最佳单模型，该单模型助力团队在500多参数队伍中取得第二名。Open Images Dataset V5(OIDV5)包含500个类别、173W训练图像和超过1400W个标注边框，是目前已知规模最大的目标检测公开数据集，数据集地址：[https://storage.googleapis.com/openimages/web/index.html](https://storage.googleapis.com/openimages/web/index.html)。团队在比赛中的技术方案报告地址：[https://arxiv.org/pdf/1911.07171.pdf](https://arxiv.org/pdf/1911.07171.pdf)
+
+<div align="center">
+  <img src="../demo/oidv5_gt.png"/>
+</div>
+
+## 方法描述
+该模型结合了当前较优的检测方法。具体地，它将ResNet200-vd作为检测模型的骨干网络，其imagenet分类预训练模型可以在[这里](https://github.com/PaddlePaddle/models/blob/develop/PaddleCV/image_classification/README_en.md)下载；结合了CascadeCA RCNN、Feature Pyramid Networks、Non-local、Deformable V2等方法。在这里需要注意的是，标准的CascadeRCNN是只预测2个框（前景和背景，使用得分信息去判断最终前景所属的类别），而该模型对每个类别都单独预测了一个框（Cascade Class Aware）。最终模型框图如下图所示。
+
+<div align="center">
+  <img src="../demo/oidv5_model_framework.png"/>
+</div>
+
+
+由于OIDV5的类别不均衡现象比较严重，在训练时采用了动态采样的策略去选择样本并进行训练；多尺度训练被用于解决边框面积范围太大的情况；此外，团队使用Libra loss替代Smooth L1 loss，来计算预测框的loss；在预测时，使用SoftNMS方法进行后处理，保证更多的框可以被召回。
+
+Objects365 Dataset和OIDV5有大约189个类别是重复的，因此将两个数据集合并进行训练，用于扩充OIDV5的训练数据，最终该模型与其性能指标如下表所示。更具体的模型训练和融合策略可以见：[OIDV5技术报告](https://arxiv.org/pdf/1911.07171.pdf)。
+
+OIDV5模型训练结果如下。
+
+
+|        模型结构         | Public/Private Score |                           下载链接                           |
+| :-----------------: | :--------: | :----------------------------------------------------------: |
+| CascadeCARCNN-FPN-Dcnv2-Nonlocal ResNet200-vd |    0.62690/0.59459    | [模型](https://paddlemodels.bj.bcebos.com/object_detection/oidv5_cascade_rcnn_cls_aware_r200_vd_fpn_dcnv2_nonlocal_softnms.tar) |
+
+
+此外，为验证模型的性能，团队基于该模型结构，也训练了针对COCO2017和Objects365 Dataset的模型，模型和验证集指标如下表。
+
+|        模型结构         | 数据集 |  验证集mAP |                           下载链接                           |
+| :-----------------: | :--------: | :--------: | :----------------------------------------------------------: |
+| CascadeCARCNN-FPN-Dcnv2-Nonlocal ResNet200-vd | COCO2017 |    51.7%    | [模型](https://paddlemodels.bj.bcebos.com/object_detection/cascade_rcnn_cls_aware_r200_vd_fpn_dcnv2_nonlocal_softnms.tar) |
+| CascadeCARCNN-FPN-Dcnv2-Nonlocal ResNet200-vd | Objects365 |    34.5%    | [模型](https://paddlemodels.bj.bcebos.com/object_detection/obj365_cascade_rcnn_cls_aware_r200_vd_fpn_dcnv2_nonlocal_softnms.tar) |
+
+COCO和Objects365 Dataset数据格式相同，目前只支持预测和评估。
+
+## 使用方法
+
+OIDV5数据集格式与COCO不同，目前仅支持单张图片的预测。OIDV5的模型评估方法可以参考[这里](https://github.com/tensorflow/models/blob/master/research/object_detection/g3doc/challenge_evaluation.md)
+
+1. 下载模型并解压。
+
+2. 运行预测程序。
+
+```
+python -u tools/infer.py -c configs/oidv5/cascade_rcnn_cls_aware_r200_vd_fpn_dcnv2_nonlocal_softnms.yml -o weights=./oidv5_cascade_rcnn_cls_aware_r200_vd_fpn_dcnv2_nonlocal_softnms/ --infer_img=demo/000000570688.jpg
+```
+
+其中模型所在文件夹需要根据自己放置的位置进行修改。
+
+检测结果图像可以在`output`文件夹中查看。
+
+## 模型检测效果
+
+<div align="center">
+  <img src="../demo/oidv5_pred.jpg"/>
+</div>

ppdet/modeling/target_assigners.py

@@ -70,6 +70,6 @@ class CascadeBBoxAssigner(object):
             bbox_reg_weights=curr_bbox_reg_w,
             use_random=self.use_random,
             class_nums=self.class_nums if self.class_aware else 2,
-            is_cls_agnostic=True,
+            is_cls_agnostic=not self.class_aware,
             is_cascade_rcnn=True if curr_stage > 0 and not self.class_aware else False)
         return outs

ppdet/utils/oid_eval.py

+# Copyright (c) 2019 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from __future__ import absolute_import
+from __future__ import division
+from __future__ import print_function
+from __future__ import unicode_literals
+
+import os
+import sys
+import numpy as np
+
+from .coco_eval import bbox2out
+
+import logging
+logger = logging.getLogger(__name__)
+
+
+__all__ = ['bbox2out', 'get_category_info']
+
+def get_category_info(anno_file=None,
+                      with_background=True,
+                      use_default_label=False):
+    clsid2catid = { k:k for k in range(1, 501) }
+
+    catid2name = {
+            0:"background",
+            1:"Infant bed",
+            2:"Rose",
+            3:"Flag",
+            4:"Flashlight",
+            5:"Sea turtle",
+            6:"Camera",
+            7:"Animal",
+            8:"Glove",
+            9:"Crocodile",
+            10:"Cattle",
+            11:"House",
+            12:"Guacamole",
+            13:"Penguin",
+            14:"Vehicle registration plate",
+            15:"Bench",
+            16:"Ladybug",
+            17:"Human nose",
+            18:"Watermelon",
+            19:"Flute",
+            20:"Butterfly",
+            21:"Washing machine",
+            22:"Raccoon",
+            23:"Segway",
+            24:"Taco",
+            25:"Jellyfish",
+            26:"Cake",
+            27:"Pen",
+            28:"Cannon",
+            29:"Bread",
+            30:"Tree",
+            31:"Shellfish",
+            32:"Bed",
+            33:"Hamster",
+            34:"Hat",
+            35:"Toaster",
+            36:"Sombrero",
+            37:"Tiara",
+            38:"Bowl",
+            39:"Dragonfly",
+            40:"Moths and butterflies",
+            41:"Antelope",
+            42:"Vegetable",
+            43:"Torch",
+            44:"Building",
+            45:"Power plugs and sockets",
+            46:"Blender",
+            47:"Billiard table",
+            48:"Cutting board",
+            49:"Bronze sculpture",
+            50:"Turtle",
+            51:"Broccoli",
+            52:"Tiger",
+            53:"Mirror",
+            54:"Bear",
+            55:"Zucchini",
+            56:"Dress",
+            57:"Volleyball",
+            58:"Guitar",
+            59:"Reptile",
+            60:"Golf cart",
+            61:"Tart",
+            62:"Fedora",
+            63:"Carnivore",
+            64:"Car",
+            65:"Lighthouse",
+            66:"Coffeemaker",
+            67:"Food processor",
+            68:"Truck",
+            69:"Bookcase",
+            70:"Surfboard",
+            71:"Footwear",
+            72:"Bench",
+            73:"Necklace",
+            74:"Flower",
+            75:"Radish",
+            76:"Marine mammal",
+            77:"Frying pan",
+            78:"Tap",
+            79:"Peach",
+            80:"Knife",
+            81:"Handbag",
+            82:"Laptop",
+            83:"Tent",
+            84:"Ambulance",
+            85:"Christmas tree",
+            86:"Eagle",
+            87:"Limousine",
+            88:"Kitchen & dining room table",
+            89:"Polar bear",
+            90:"Tower",
+            91:"Football",
+            92:"Willow",
+            93:"Human head",
+            94:"Stop sign",
+            95:"Banana",
+            96:"Mixer",
+            97:"Binoculars",
+            98:"Dessert",
+            99:"Bee",
+            100:"Chair",
+            101:"Wood-burning stove",
+            102:"Flowerpot",
+            103:"Beaker",
+            104:"Oyster",
+            105:"Woodpecker",
+            106:"Harp",
+            107:"Bathtub",
+            108:"Wall clock",
+            109:"Sports uniform",
+            110:"Rhinoceros",
+            111:"Beehive",
+            112:"Cupboard",
+            113:"Chicken",
+            114:"Man",
+            115:"Blue jay",
+            116:"Cucumber",
+            117:"Balloon",
+            118:"Kite",
+            119:"Fireplace",
+            120:"Lantern",
+            121:"Missile",
+            122:"Book",
+            123:"Spoon",
+            124:"Grapefruit",
+            125:"Squirrel",
+            126:"Orange",
+            127:"Coat",
+            128:"Punching bag",
+            129:"Zebra",
+            130:"Billboard",
+            131:"Bicycle",
+            132:"Door handle",
+            133:"Mechanical fan",
+            134:"Ring binder",
+            135:"Table",
+            136:"Parrot",
+            137:"Sock",
+            138:"Vase",
+            139:"Weapon",
+            140:"Shotgun",
+            141:"Glasses",
+            142:"Seahorse",
+            143:"Belt",
+            144:"Watercraft",
+            145:"Window",
+            146:"Giraffe",
+            147:"Lion",
+            148:"Tire",
+            149:"Vehicle",
+            150:"Canoe",
+            151:"Tie",
+            152:"Shelf",
+            153:"Picture frame",
+            154:"Printer",
+            155:"Human leg",
+            156:"Boat",
+            157:"Slow cooker",
+            158:"Croissant",
+            159:"Candle",
+            160:"Pancake",
+            161:"Pillow",
+            162:"Coin",
+            163:"Stretcher",
+            164:"Sandal",
+            165:"Woman",
+            166:"Stairs",
+            167:"Harpsichord",
+            168:"Stool",
+            169:"Bus",
+            170:"Suitcase",
+            171:"Human mouth",
+            172:"Juice",
+            173:"Skull",
+            174:"Door",
+            175:"Violin",
+            176:"Chopsticks",
+            177:"Digital clock",
+            178:"Sunflower",
+            179:"Leopard",
+            180:"Bell pepper",
+            181:"Harbor seal",
+            182:"Snake",
+            183:"Sewing machine",
+            184:"Goose",
+            185:"Helicopter",
+            186:"Seat belt",
+            187:"Coffee cup",
+            188:"Microwave oven",
+            189:"Hot dog",
+            190:"Countertop",
+            191:"Serving tray",
+            192:"Dog bed",
+            193:"Beer",
+            194:"Sunglasses",
+            195:"Golf ball",
+            196:"Waffle",
+            197:"Palm tree",
+            198:"Trumpet",
+            199:"Ruler",
+            200:"Helmet",
+            201:"Ladder",
+            202:"Office building",
+            203:"Tablet computer",
+            204:"Toilet paper",
+            205:"Pomegranate",
+            206:"Skirt",
+            207:"Gas stove",
+            208:"Cookie",
+            209:"Cart",
+            210:"Raven",
+            211:"Egg",
+            212:"Burrito",
+            213:"Goat",
+            214:"Kitchen knife",
+            215:"Skateboard",
+            216:"Salt and pepper shakers",
+            217:"Lynx",
+            218:"Boot",
+            219:"Platter",
+            220:"Ski",
+            221:"Swimwear",
+            222:"Swimming pool",
+            223:"Drinking straw",
+            224:"Wrench",
+            225:"Drum",
+            226:"Ant",
+            227:"Human ear",
+            228:"Headphones",
+            229:"Fountain",
+            230:"Bird",
+            231:"Jeans",
+            232:"Television",
+            233:"Crab",
+            234:"Microphone",
+            235:"Home appliance",
+            236:"Snowplow",
+            237:"Beetle",
+            238:"Artichoke",
+            239:"Jet ski",
+            240:"Stationary bicycle",
+            241:"Human hair",
+            242:"Brown bear",
+            243:"Starfish",
+            244:"Fork",
+            245:"Lobster",
+            246:"Corded phone",
+            247:"Drink",
+            248:"Saucer",
+            249:"Carrot",
+            250:"Insect",
+            251:"Clock",
+            252:"Castle",
+            253:"Tennis racket",
+            254:"Ceiling fan",
+            255:"Asparagus",
+            256:"Jaguar",
+            257:"Musical instrument",
+            258:"Train",
+            259:"Cat",
+            260:"Rifle",
+            261:"Dumbbell",
+            262:"Mobile phone",
+            263:"Taxi",
+            264:"Shower",
+            265:"Pitcher",
+            266:"Lemon",
+            267:"Invertebrate",
+            268:"Turkey",
+            269:"High heels",
+            270:"Bust",
+            271:"Elephant",
+            272:"Scarf",
+            273:"Barrel",
+            274:"Trombone",
+            275:"Pumpkin",
+            276:"Box",
+            277:"Tomato",
+            278:"Frog",
+            279:"Bidet",
+            280:"Human face",
+            281:"Houseplant",
+            282:"Van",
+            283:"Shark",
+            284:"Ice cream",
+            285:"Swim cap",
+            286:"Falcon",
+            287:"Ostrich",
+            288:"Handgun",
+            289:"Whiteboard",
+            290:"Lizard",
+            291:"Pasta",
+            292:"Snowmobile",
+            293:"Light bulb",
+            294:"Window blind",
+            295:"Muffin",
+            296:"Pretzel",
+            297:"Computer monitor",
+            298:"Horn",
+            299:"Furniture",
+            300:"Sandwich",
+            301:"Fox",
+            302:"Convenience store",
+            303:"Fish",
+            304:"Fruit",
+            305:"Earrings",
+            306:"Curtain",
+            307:"Grape",
+            308:"Sofa bed",
+            309:"Horse",
+            310:"Luggage and bags",
+            311:"Desk",
+            312:"Crutch",
+            313:"Bicycle helmet",
+            314:"Tick",
+            315:"Airplane",
+            316:"Canary",
+            317:"Spatula",
+            318:"Watch",
+            319:"Lily",
+            320:"Kitchen appliance",
+            321:"Filing cabinet",
+            322:"Aircraft",
+            323:"Cake stand",
+            324:"Candy",
+            325:"Sink",
+            326:"Mouse",
+            327:"Wine",
+            328:"Wheelchair",
+            329:"Goldfish",
+            330:"Refrigerator",
+            331:"French fries",
+            332:"Drawer",
+            333:"Treadmill",
+            334:"Picnic basket",
+            335:"Dice",
+            336:"Cabbage",
+            337:"Football helmet",
+            338:"Pig",
+            339:"Person",
+            340:"Shorts",
+            341:"Gondola",
+            342:"Honeycomb",
+            343:"Doughnut",
+            344:"Chest of drawers",
+            345:"Land vehicle",
+            346:"Bat",
+            347:"Monkey",
+            348:"Dagger",
+            349:"Tableware",
+            350:"Human foot",
+            351:"Mug",
+            352:"Alarm clock",
+            353:"Pressure cooker",
+            354:"Human hand",
+            355:"Tortoise",
+            356:"Baseball glove",
+            357:"Sword",
+            358:"Pear",
+            359:"Miniskirt",
+            360:"Traffic sign",
+            361:"Girl",
+            362:"Roller skates",
+            363:"Dinosaur",
+            364:"Porch",
+            365:"Human beard",
+            366:"Submarine sandwich",
+            367:"Screwdriver",
+            368:"Strawberry",
+            369:"Wine glass",
+            370:"Seafood",
+            371:"Racket",
+            372:"Wheel",
+            373:"Sea lion",
+            374:"Toy",
+            375:"Tea",
+            376:"Tennis ball",
+            377:"Waste container",
+            378:"Mule",
+            379:"Cricket ball",
+            380:"Pineapple",
+            381:"Coconut",
+            382:"Doll",
+            383:"Coffee table",
+            384:"Snowman",
+            385:"Lavender",
+            386:"Shrimp",
+            387:"Maple",
+            388:"Cowboy hat",
+            389:"Goggles",
+            390:"Rugby ball",
+            391:"Caterpillar",
+            392:"Poster",
+            393:"Rocket",
+            394:"Organ",
+            395:"Saxophone",
+            396:"Traffic light",
+            397:"Cocktail",
+            398:"Plastic bag",
+            399:"Squash",
+            400:"Mushroom",
+            401:"Hamburger",
+            402:"Light switch",
+            403:"Parachute",
+            404:"Teddy bear",
+            405:"Winter melon",
+            406:"Deer",
+            407:"Musical keyboard",
+            408:"Plumbing fixture",
+            409:"Scoreboard",
+            410:"Baseball bat",
+            411:"Envelope",
+            412:"Adhesive tape",
+            413:"Briefcase",
+            414:"Paddle",
+            415:"Bow and arrow",
+            416:"Telephone",
+            417:"Sheep",
+            418:"Jacket",
+            419:"Boy",
+            420:"Pizza",
+            421:"Otter",
+            422:"Office supplies",
+            423:"Couch",
+            424:"Cello",
+            425:"Bull",
+            426:"Camel",
+            427:"Ball",
+            428:"Duck",
+            429:"Whale",
+            430:"Shirt",
+            431:"Tank",
+            432:"Motorcycle",
+            433:"Accordion",
+            434:"Owl",
+            435:"Porcupine",
+            436:"Sun hat",
+            437:"Nail",
+            438:"Scissors",
+            439:"Swan",
+            440:"Lamp",
+            441:"Crown",
+            442:"Piano",
+            443:"Sculpture",
+            444:"Cheetah",
+            445:"Oboe",
+            446:"Tin can",
+            447:"Mango",
+            448:"Tripod",
+            449:"Oven",
+            450:"Mouse",
+            451:"Barge",
+            452:"Coffee",
+            453:"Snowboard",
+            454:"Common fig",
+            455:"Salad",
+            456:"Marine invertebrates",
+            457:"Umbrella",
+            458:"Kangaroo",
+            459:"Human arm",
+            460:"Measuring cup",
+            461:"Snail",
+            462:"Loveseat",
+            463:"Suit",
+            464:"Teapot",
+            465:"Bottle",
+            466:"Alpaca",
+            467:"Kettle",
+            468:"Trousers",
+            469:"Popcorn",
+            470:"Centipede",
+            471:"Spider",
+            472:"Sparrow",
+            473:"Plate",
+            474:"Bagel",
+            475:"Personal care",
+            476:"Apple",
+            477:"Brassiere",
+            478:"Bathroom cabinet",
+            479:"studio couch",
+            480:"Computer keyboard",
+            481:"Table tennis racket",
+            482:"Sushi",
+            483:"Cabinetry",
+            484:"Street light",
+            485:"Towel",
+            486:"Nightstand",
+            487:"Rabbit",
+            488:"Dolphin",
+            489:"Dog",
+            490:"Jug",
+            491:"Wok",
+            492:"Fire hydrant",
+            493:"Human eye",
+            494:"Skyscraper",
+            495:"Backpack",
+            496:"Potato",
+            497:"Paper towel",
+            498:"Lifejacket",
+            499:"Bicycle wheel",
+            500:"Toilet",
+    }
+    
+    if not with_background:
+        clsid2catid = {k - 1: v for k, v in clsid2catid.items()}
+    return clsid2catid, catid2name
+    
\ No newline at end of file

tools/eval.py

@@ -106,13 +106,16 @@ def main():
     exe.run(startup_prog)
     if 'weights' in cfg:
         checkpoint.load_params(exe, eval_prog, cfg.weights)
-
+    
+    assert cfg.metric != 'OID', "eval process of OID dataset \
+                          is not supported."
     if cfg.metric == "WIDERFACE":
         raise ValueError("metric type {} does not support in tools/eval.py, "
                          "please use tools/face_eval.py".format(cfg.metric))
     assert cfg.metric in ['COCO', 'VOC'], \
             "unknown metric type {}".format(cfg.metric)
     extra_keys = []
+    
     if cfg.metric == 'COCO':
         extra_keys = ['im_info', 'im_id', 'im_shape']
     if cfg.metric == 'VOC':

tools/infer.py

@@ -139,10 +139,10 @@ def main():
         checkpoint.load_params(exe, infer_prog, cfg.weights)
 
     # parse infer fetches
-    assert cfg.metric in ['COCO', 'VOC', 'WIDERFACE'], \
+    assert cfg.metric in ['COCO', 'VOC', 'OID', 'WIDERFACE'], \
             "unknown metric type {}".format(cfg.metric)
     extra_keys = []
-    if cfg['metric'] == 'COCO':
+    if cfg['metric'] in ['COCO', 'OID']:
         extra_keys = ['im_info', 'im_id', 'im_shape']
     if cfg['metric'] == 'VOC' or cfg['metric'] == 'WIDERFACE':
         extra_keys = ['im_id', 'im_shape']
@@ -151,6 +151,8 @@ def main():
     # parse dataset category
     if cfg.metric == 'COCO':
         from ppdet.utils.coco_eval import bbox2out, mask2out, get_category_info
+    if cfg.metric == 'OID':
+        from ppdet.utils.oid_eval import bbox2out, get_category_info
     if cfg.metric == "VOC":
         from ppdet.utils.voc_eval import bbox2out, get_category_info
     if cfg.metric == "WIDERFACE":