Skip to content

Releases: breezedeus/CnSTD

fix device value

02 Jul 12:52
259e8ba
Compare
Choose a tag to compare
Merge pull request #61 from breezedeus/pytorch

fix device value

fix: clip the results

01 Jul 15:20
40dd181
Compare
Choose a tag to compare
Merge pull request #60 from breezedeus/pytorch

fix: clip the results

use hf_hub_download to download model files from huggingface

30 Jun 04:40
d46df69
Compare
Choose a tag to compare
Merge pull request #59 from breezedeus/pytorch

use hf_hub_download to download model files from huggingface

support yolov7 base model for mfd

19 Feb 03:35
a9e1c11
Compare
Choose a tag to compare

主要变更:

  • MFD训练了参数更多精度更高的模型,供 P2T网页版 使用。
  • 优化了检测出的boxes的排序算法,使得boxes的顺序更加符合人类的阅读习惯。

add MFD and Layout Analysis models

01 Feb 06:41
9da5b98
Compare
Choose a tag to compare

add MFD and Layout Analysis models

V1.2: 加入了对 PaddleOCR 检测模型的支持

06 Jul 13:00
15fa3d1
Compare
Choose a tag to compare

主要变更:

  • 加入了对 PaddleOCR 检测模型的支持;
  • 部分调整了检测结果中 box 的表达方式,统一为 4 个点的坐标值;
  • 修复了已知bugs。

fix: for opencv-python >=4.5.2

27 May 14:00
01c5a50
Compare
Choose a tag to compare

for opencv-python >=4.5.2

  • 修复图片反转问题

v1.1.1: Merge pull request #43 from breezedeus/pytorch

27 May 08:27
0e697c0
Compare
Choose a tag to compare

修复训练bug,重新训练获得更优模型;提供了更多更小的预训练模型

23 Sep 14:26
Compare
Choose a tag to compare

修复对 torchvision 依赖版本要求

30 Aug 05:38
cecb047
Compare
Choose a tag to compare
Merge pull request #29 from breezedeus/pytorch

fix torchvision version dependency