From bec64c2296424ad82e8f27ea8e3959d6ab20bb47 Mon Sep 17 00:00:00 2001
From: FlyingQianMM <245467267@qq.com>
Date: Thu, 4 Jun 2020 14:02:01 +0000
Subject: [PATCH] delete recall drop

---
 docs/tuning_strategy/detection/negatives_training.md | 1 -
 1 file changed, 1 deletion(-)

diff --git a/docs/tuning_strategy/detection/negatives_training.md b/docs/tuning_strategy/detection/negatives_training.md
index 5855010..63cd54a 100644
--- a/docs/tuning_strategy/detection/negatives_training.md
+++ b/docs/tuning_strategy/detection/negatives_training.md
@@ -8,7 +8,6 @@
 
 * 与基准模型相比，通过负样本学习后的模型**mmAP有3.6%的提升，mAP有0.1%的提升**。
 * 与基准模型相比，通过负样本学习后的模型在背景图片上的图片级别**误检率降低了49.68%**。
-* 与基准模型相比，通过负样本学习后的模型在目标图片上的图片级别**召回率仅降低了1.22%**。
 
 表1 违禁品验证集上**框级别精度**对比
 
-- 
GitLab