Asunany 于 2018-01-26 21:35:40 发布
博主在使用Faster R-CNN进行目标检测时遇到问题，训练后得到的AP和mAP均为0。环境是ubuntu + CUDA + Python，使用VOC2007数据集训练模型。尽管模型可以运行，但检测精度为0，可能的原因在代码中出现警告，提示有无效值。博主寻求社区帮助解决这个问题。 大家好,本人最近在学习faster rcnn ,小白一枚,参考了很多教程。
环境: ubuntu+cuda+python
运行结果:
Saving cached annotations to /home/think/asuna/py-faster-rcnn/data/VOCdevkit2007/annotations_cache/annots.pkl
/home/think/asuna/py-faster-rcnn/tools/../lib/datasets/voc_eval.py:196: RuntimeWarning: invalid value encountered in divide
rec = tp / float(npos)
/home/think/asuna/py-faster-rcnn/tools/../lib/datasets/voc_eval.py:41: RuntimeWarning: invalid value encountered in greater_equal
if np.sum(rec >= t) == 0:
AP for face = 0.0000
Mean AP = 0.0000
~~~~~~~~
Results:
0.000
0.000
~~~~~~~~
--------------------------------------------------------------
Results computed with the **unofficial** Python eval code.
Results should be very close to the official MATLAB eval code.
Recompute with `./tools/reval.py --matlab ...` for
