DongYongFei comments

Results 7 comments of


                                            DongYongFei

为什么我产生的拓扑图和作者产生的不一样，而且差别还挺大的

![image](https://user-images.githubusercontent.com/54470509/111136253-7ea2b080-85b8-11eb-82b5-9c5d660e6189.png) 这是我生成的yolov5s.xml ![image](https://user-images.githubusercontent.com/54470509/111136463-bd386b00-85b8-11eb-82ca-e39d32728965.png) 这是作者的yolov5s.xml 显然，拓扑结构是不一样的我运行我产生的yolov5s.xml结果 ![image](https://user-images.githubusercontent.com/54470509/111136597-e0fbb100-85b8-11eb-8933-26307f8929d2.png) 很奇怪，我也是v3.1

If I want to recognize a picture, how can I do it? Do I still want to generate LMDB format? Can you provide a predict interface? Thank you

Can you provide a simple demo？thank you

evaluate_with_lexicon

okay, thank you for your outstanding work

基于PaddleInference的预测 SER + RE模型串联报错 - Hint: Expected dtype() == paddle::experimental::CppTypeToDataType<T>::Type(), but received dtype():5 != paddle::experimental::CppTypeToDataType<T>::Type():7.] (at C:\home\workspace\Paddle_release\paddle\phi\core\dense_tensor.cc:137) [operator < less_equal > error

> > python3 predict_system.py > > --kie_algorithm=LayoutXLM > > --re_model_dir=./inference/re_vi_layoutxlm_xfund_infer > > --ser_model_dir=./inference/ser_vi_layoutxlm_xfund_infer > > --image_dir=./docs/kie/input/zh_val_42.jpg > > --ser_dict_path=../ppocr/utils/dict/kie_dict/xfund_class_list.txt > > --vis_font_path=../doc/fonts/simfang.ttf > > --ocr_order_method="tb-yx" > > --mode=kie > >...

DongYongFei

为什么我产生的拓扑图和作者产生的不一样，而且差别还挺大的

If I want to recognize a picture, how can I do it? Do I still want to generate LMDB format? Can you provide a predict interface? Thank you

evaluate_with_lexicon

layoutlmv2进行ser+re联合推理出错（单独ser推理正常），权重均来自于官方。layoutxlm和vi_layoutxlm均能正常推理。

Are the parameters of swinTransformer trained during fine-tuning?

Training on FUNSD: Cuda out of Memory on GPU with 12Gb memory.