From e7ba2d852bfd425ed05e5b472570cc4650817c92 Mon Sep 17 00:00:00 2001 From: dengkaipeng Date: Wed, 17 Apr 2019 06:47:47 +0000 Subject: [PATCH] add score filter doc. --- PaddleCV/yolov3/README.md | 2 ++ PaddleCV/yolov3/README_cn.md | 2 ++ 2 files changed, 4 insertions(+) diff --git a/PaddleCV/yolov3/README.md b/PaddleCV/yolov3/README.md index a079257e..67055130 100644 --- a/PaddleCV/yolov3/README.md +++ b/PaddleCV/yolov3/README.md @@ -142,6 +142,8 @@ Evalutaion result is shown as below: | 416x416 | 36.5 | 58.2 | 39.1 | | 320x320 | 34.1 | 55.4 | 36.3 | +- **NOTE:** evaluations based on `pycocotools` evaluator, predict bounding boxes with `score < 0.05` were not filtered out. Some frameworks which filtered out predict bounding boxes with `score < 0.05` will cause a drop in accuracy. + ## Inference and Visualization Inference is used to get prediction score or image features based on trained models. `infer.py` is the main executor for inference, one can start infer step by: diff --git a/PaddleCV/yolov3/README_cn.md b/PaddleCV/yolov3/README_cn.md index 79d62745..9ad33c67 100644 --- a/PaddleCV/yolov3/README_cn.md +++ b/PaddleCV/yolov3/README_cn.md @@ -142,6 +142,8 @@ Train Loss | 416x416 | 36.5 | 58.2 | 39.1 | | 320x320 | 34.1 | 55.4 | 36.3 | +- **注意: **评估结果基于`pycocotools`评估器,没有滤除`score < 0.05`的预测框,其他框架有此滤除操作会导致精度下降。 + ## 模型推断及可视化 -- GitLab