SwinTextSpotter icon indicating copy to clipboard operation
SwinTextSpotter copied to clipboard

Pytorch re-implementation of Paper: SwinTextSpotter: Scene Text Spotting via Better Synergy between Text Detection and Text Recognition (CVPR 2022)

Results 56 SwinTextSpotter issues
Sort by recently updated
recently updated
newest added

您好,执行`python setup.py build develop`过程中出现如下报错,要如何解决?谢谢! Traceback (most recent call last): File "/home/zyq/anaconda3/envs/SWINTS/lib/python3.8/site-packages/pkg_resources/__init__.py", line 3109, in _dep_map return self.__dep_map File "/home/zyq/anaconda3/envs/SWINTS/lib/python3.8/site-packages/pkg_resources/__init__.py", line 2902, in __getattr__ raise AttributeError(attr) AttributeError: _DistInfoDistribution__dep_map During handling of...

下载数据那里的tataltext的json文件是怎么生成的呀/(ㄒoㄒ)/~~

刚开始接触这个领域 我想请教一下 我现在只有原图 和 每一张图对应的txt(txt里面包括了边界框的四个坐标和对应的识别字符) 我应该如何生成对应的json文件并进行训练呢

从https://github.com/mxin262/SwinTextSpotter/blob/main/chn_cls_list.txt下载的文件乱码 ![1](https://github.com/mxin262/SwinTextSpotter/assets/110005548/58f5b9d8-c244-4a24-ac8c-ff34275f2a65)

作者你好,我是了解到ABCNet后查找到最新的SwinTextSpotter项目,我觉得它应该是比ABCNet更优秀,您在SwinTextSpotter提到更新的ESTextSpotter,我在昨天尝试之后,能 执行 vis.py (需在for循环末尾添加: torch.cuda.empty_cache(),) 我的GPU都是独立的,3060,3070,4090。单张GPU资源能否支持这两个项目之一,我的资源太少了,没有8张。我即使GPU跑一张图显存也不够(在3060实验ESTextSpotter,还未测试4090,)。这两项目对资源的需求能否再下降? 这是可行的吗? 如果我对网络不管精度先剪枝确保能训练这是能被推荐的吗? 还是我这边添加设备会更好些?

REC_HEAD: BATCH_SIZE: 128 SOLVER: IMS_PER_BATCH: 8 IMS_PER_BATCH是什么

请问一下,我想训练自己的数据集,该怎么着手呢?

pkg_resources.extern.packaging.requirements.InvalidRequirement: .* suffix can only be used with `==` or `!=` operators PyYAML (>=5.1.*) ~~~~~~^

Hello! We conducted training with the code below using the totaltext dataset. But score is like this "E2E_RESULTS: precision: 0, recall: 0.0, hmean: 0" after training. And as a result...