Strange evaluation output when evaluating

Describe the bug Hi guys, I saw some strange evaluation output when evaluating, as shown below,

+------------+-----+------+--------+-------+
| class      | gts | dets | recall | ap    |
+------------+-----+------+--------+-------+
| helicopter | 77  | 106  | 1.000  | 1.000 |
| c130       | 318 | 362  | 0.997  | 0.997 |
| b2         | 6   | 15   | 1.000  | 1.000 |
| other      | 76  | 145  | 0.961  | 0.946 |
| b52        | 73  | 142  | 0.986  | 0.986 |
| kc10       | 43  | 95   | 1.000  | 0.999 |
| c17        | 102 | 194  | 1.000  | 0.992 |
| command    | 10  | 66   | 1.000  | 0.961 |
| c5         | 27  | 139  | 1.000  | 0.937 |
| f15        | 75  | 196  | 1.000  | 0.998 |
| kc135      | 130 | 237  | 1.000  | 0.998 |
| a10        | 22  | 45   | 1.000  | 0.998 |
| b1         | 78  | 133  | 1.000  | 0.998 |
| aew        | 55  | 80   | 1.000  | 0.999 |
| p3         | 49  | 73   | 1.000  | 1.000 |
| f16        | 135 | 271  | 0.985  | 0.983 |
| p8         | 24  | 42   | 1.000  | 0.985 |
| f35        | 27  | 95   | 1.000  | 0.985 |
| f18        | 116 | 175  | 1.000  | 0.999 |
| v22        | 56  | 68   | 0.982  | 0.982 |
| f4         | 4   | 65   | 1.000  | 0.900 |
| f22        | 21  | 72   | 1.000  | 0.996 |
| globalhawk | 13  | 37   | 1.000  | 0.984 |
| u2         | 4   | 9    | 1.000  | 1.000 |
+------------+-----+------+--------+-------+
| mAP        |     |      |        | 0.984 |
+------------+-----+------+--------+-------+

It is strange that the AP would reach 1, which seems impossible in a real scenario.

Reproduction

1. What command or script did you run?

python tools/voc_eval.py results.pkl configs/test_faster_rcnn_r50_fpn_1x_voc0712.py

1. What dataset did you use? I use my own dataset in VOC format.

Environment

1. Please run python tools/collect_env.py to collect necessary environment infomation and paste it here.

sys.platform: linux
Python: 3.7.4 (default, Aug 13 2019, 20:35:49) [GCC 7.3.0]
CUDA available: True
CUDA_HOME: /usr/local/cuda
NVCC: Cuda compilation tools, release 10.1, V10.1.243
GPU 0: GeForce RTX 2080 Ti
GCC: gcc (Ubuntu 7.4.0-1ubuntu1~18.04.1) 7.4.0
PyTorch: 1.3.1
PyTorch compiling details: PyTorch built with:
  - GCC 7.3
  - Intel(R) Math Kernel Library Version 2019.0.4 Product Build 20190411 for Intel(R) 64 architecture applications
  - Intel(R) MKL-DNN v0.20.5 (Git Hash 0125f28c61c1f822fd48570b4c1066f96fcb9b2e)
  - OpenMP 201511 (a.k.a. OpenMP 4.5)
  - NNPACK is enabled
  - CUDA Runtime 10.1
  - NVCC architecture flags: -gencode;arch=compute_37,code=sm_37;-gencode;arch=compute_50,code=sm_50;-gencode;arch=compute_60,code=sm_60;-gencode;arch=compute_61,code=sm_61;-gencode;arch=compute_70,code=sm_70;-gencode;arch=compute_75,code=sm_75;-gencode;arch=compute_37,code=compute_37
  - CuDNN 7.6.3
  - Magma 2.5.1
  - Build settings: BLAS=MKL, BUILD_NAMEDTENSOR=OFF, BUILD_TYPE=Release, CXX_FLAGS= -Wno-deprecated -fvisibility-inlines-hidden -fopenmp -DUSE_FBGEMM -DUSE_QNNPACK -DUSE_PYTORCH_QNNPACK -O2 -fPIC -Wno-narrowing -Wall -Wextra -Wno-missing-field-initializers -Wno-type-limits -Wno-array-bounds -Wno-unknown-pragmas -Wno-sign-compare -Wno-unused-parameter -Wno-unused-variable -Wno-unused-function -Wno-unused-result -Wno-strict-overflow -Wno-strict-aliasing -Wno-error=deprecated-declarations -Wno-stringop-overflow -Wno-error=pedantic -Wno-error=redundant-decls -Wno-error=old-style-cast -fdiagnostics-color=always -faligned-new -Wno-unused-but-set-variable -Wno-maybe-uninitialized -fno-math-errno -fno-trapping-math -Wno-stringop-overflow, DISABLE_NUMA=1, PERF_WITH_AVX=1, PERF_WITH_AVX2=1, PERF_WITH_AVX512=1, USE_CUDA=True, USE_EXCEPTION_PTR=1, USE_GFLAGS=OFF, USE_GLOG=OFF, USE_MKL=ON, USE_MKLDNN=ON, USE_MPI=OFF, USE_NCCL=ON, USE_NNPACK=ON, USE_OPENMP=ON, USE_STATIC_DISPATCH=OFF, 

TorchVision: 0.4.2
OpenCV: 4.1.1
MMCV: 0.2.15
MMDetection: 1.0rc1+e907139
MMDetection Compiler: GCC 7.4
MMDetection CUDA Compiler: 10.1

1. You may add addition that may be helpful for locating the problem, such as
   • How you installed PyTorch [e.g., pip, conda, source] By Conda commands.

Bug fix If you have already identified the reason, you can provide the information here. If you are willing to create a PR to fix it, please also leave a comment here and that would be much appreciated!

[hellock]

Not really, it doesn’t matter if all boxes have IoU greater than 0.5. You may refer to the computation of mAP: https://github.com/open-mmlab/mmdetection/blob/master/mmdet/core/evaluation/mean_ap.py

[hellock]

We cannot help since it is unclear about your custom dataset, your training, or anything else. Any value between 0 and 1 can be a valid mAP.