Liang Qiao

Have you run the setup.sh script? It will automatically install all the dependencies including 'mmdet' and 'mmcv'.

If the mmcv cannot be intalled by PIP, you may find some solutions from the [official repository](https://github.com/open-mmlab/mmcv). I'm not sure if mmcv-full-1.4.0 is matched with mmdetection-2.11.0. mmcv-full-1.3.4 works well in...

Thank you for your suggestion. We will add more details to optimize the demos, including how to used separate modules and how can different algorithms work in a pipeline, etc.

Thank you for your suggestion. We will consider adding more details to optimize the demos, such as how to used separate modules and how can different algorithms work in a...

Thank you for your suggestion. We will consider adding more details to optimize the demos, such as how to used separate modules and how can different algorithms work in a...

这个问题看上去像是pytorch版本或其他环境变量的问题，是否使用的是gpu版本的pytorch。 如果不训练识别模型话，也可以绕过这个包的编译（把相关包的引用语句注释掉，例如import warpctc），不影响框架在其他模型上的正常运行。

> 这个问题看上去像是pytorch版本或其他环境变量的问题，是否使用的是gpu版本的pytorch。 > > 如果不训练识别模型话，也可以绕过这个包的编译（把相关包的引用语句注释掉，例如import warpctc），不影响框架在其他模型上的正常运行。 从网上上找到的资料看，可能是由于最新版本的pytorch (1.11以上）去除对这个头文件的支持，所以我们建议使用pytorch 1.8版本会比较稳定

It looks like the model is not well learned, can you check if the prediction target of the model is correct (L163 in pipelines/gpma_data.py)?

The problem comes from 'OSError: GPMADataGeneration: no file with expected extension'. You should run the setup.sh to generate the "*.so" file. If it is correctly excuted, there will be a...

It is because the algorithm of EAST and Warpctc require other dependencies. It won’t affect the model of LGPMA, You can simply ignore them by comment the related code in...