提交 0865d0c4 编写于 作者: 文幕地方's avatar 文幕地方

rm \n in FCEMetric

上级 d5363df4
# 两阶段算法 # 两阶段算法
- [两阶段算法](#-----) - [两阶段算法](#两阶段算法)
* [1. 算法介绍](#1) - [1. 算法介绍](#1-算法介绍)
+ [1.1 文本检测算法](#11) - [1.1 文本检测算法](#11-文本检测算法)
+ [1.2 文本识别算法](#12) - [1.2 文本识别算法](#12-文本识别算法)
* [2. 模型训练](#2) - [2. 模型训练](#2-模型训练)
* [3. 模型推理](#3) - [3. 模型推理](#3-模型推理)
<a name="1"></a> <a name="1"></a>
...@@ -44,7 +44,7 @@ PaddleOCR开源的文本检测算法列表: ...@@ -44,7 +44,7 @@ PaddleOCR开源的文本检测算法列表:
|模型|骨干网络|precision|recall|Hmean|下载链接| |模型|骨干网络|precision|recall|Hmean|下载链接|
| --- | --- | --- | --- | --- | --- | | --- | --- | --- | --- | --- | --- |
|FCE|ResNet50_dcn|88.13%|82.60%|85.28%|[训练模型](https://paddleocr.bj.bcebos.com/contribution/det_r50_dcn_fce_ctw_v2.0_train.tar)| |FCE|ResNet50_dcn|88.39%|82.18%|85.27%|[训练模型](https://paddleocr.bj.bcebos.com/contribution/det_r50_dcn_fce_ctw_v2.0_train.tar)|
**说明:** SAST模型训练额外加入了icdar2013、icdar2017、COCO-Text、ArT等公开数据集进行调优。PaddleOCR用到的经过整理格式的英文公开数据集下载: **说明:** SAST模型训练额外加入了icdar2013、icdar2017、COCO-Text、ArT等公开数据集进行调优。PaddleOCR用到的经过整理格式的英文公开数据集下载:
* [百度云地址](https://pan.baidu.com/s/12cPnZcVuV1zn5DOd4mqjVw) (提取码: 2bpi) * [百度云地址](https://pan.baidu.com/s/12cPnZcVuV1zn5DOd4mqjVw) (提取码: 2bpi)
......
...@@ -135,7 +135,7 @@ class DetFCEMetric(object): ...@@ -135,7 +135,7 @@ class DetFCEMetric(object):
# metircs['{}_{}'.format(key, score_thr)] = value # metircs['{}_{}'.format(key, score_thr)] = value
metirc_str = 'precision:{:.5f} recall:{:.5f} hmean:{:.5f}'.format( metirc_str = 'precision:{:.5f} recall:{:.5f} hmean:{:.5f}'.format(
metirc['precision'], metirc['recall'], metirc['hmean']) metirc['precision'], metirc['recall'], metirc['hmean'])
metircs['\n thr {}'.format(score_thr)] = metirc_str metircs['thr {}'.format(score_thr)] = metirc_str
hmean = max(hmean, metirc['hmean']) hmean = max(hmean, metirc['hmean'])
metircs['hmean'] = hmean metircs['hmean'] = hmean
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册