diff --git a/pplayout/README.md b/ppstructure/vqa/README.md similarity index 93% rename from pplayout/README.md rename to ppstructure/vqa/README.md index 2de1f5824142ad1013a46c0ff326464702a3bf0a..8d117fdeb16e1c0e90bf6ec89924e414fc764249 100644 --- a/pplayout/README.md +++ b/ppstructure/vqa/README.md @@ -1,6 +1,6 @@ -# PP-Layout +# 视觉问答(VQA) -PP-Layout是一个针对图像文本内容进行语义实体识别与关系抽取的OCR工具包,主要特性如下: +VQA主要特性如下: - 集成[LayoutXLM](https://arxiv.org/pdf/2104.08836.pdf)模型以及PP-OCR预测引擎。 - 支持基于多模态方法的语义实体识别 (Semantic Entity Recognition, SER) 以及关系抽取 (Relation Extraction, RE) 任务。基于 SER 任务,可以完成对图像中的文本识别与分类;基于 RE 任务,可以完成对图象中的文本内容的关系提取(比如判断问题对) @@ -22,7 +22,7 @@ PP-Layout是一个针对图像文本内容进行语义实体识别与关系抽
- +
其中不同颜色的框表示不同的类别,对于XFUN数据集,有`QUESTION`, `ANSWER`, `HEADER` 3种类别,在OCR检测框的左上方也标出了对应的类别和OCR识别结果。 @@ -53,7 +53,7 @@ python3 -m pip install paddlepaddle==2.2 -i https://mirror.baidu.com/pypi/simple 更多需求,请参照[安装文档](https://www.paddlepaddle.org.cn/install/quick)中的说明进行操作。 -### 2.2 安装PaddleOCR(包含PP-OCR和PP-Layout) +### 2.2 安装PaddleOCR(包含 PP-OCR 和 VQA ) - **(1)pip快速安装PaddleOCR whl包(仅预测)** @@ -61,7 +61,7 @@ python3 -m pip install paddlepaddle==2.2 -i https://mirror.baidu.com/pypi/simple pip install "paddleocr>=2.2" # 推荐使用2.2+版本 ``` -- **(2)下载PP-Layout源码(预测+训练)** +- **(2)下载VQA源码(预测+训练)** ```bash 【推荐】git clone https://github.com/PaddlePaddle/PaddleOCR @@ -82,7 +82,7 @@ pip install -e . ``` -- **(4)安装PP-Layout的`requirements`** +- **(4)安装VQA的`requirements`** ```bash pip install -r requirements.txt diff --git a/pplayout/helper/eval_with_label_end2end.py b/ppstructure/vqa/helper/eval_with_label_end2end.py similarity index 100% rename from pplayout/helper/eval_with_label_end2end.py rename to ppstructure/vqa/helper/eval_with_label_end2end.py diff --git a/pplayout/helper/trans_xfun_data.py b/ppstructure/vqa/helper/trans_xfun_data.py similarity index 100% rename from pplayout/helper/trans_xfun_data.py rename to ppstructure/vqa/helper/trans_xfun_data.py diff --git a/pplayout/images/input/zh_val_0.jpg b/ppstructure/vqa/images/input/zh_val_0.jpg similarity index 100% rename from pplayout/images/input/zh_val_0.jpg rename to ppstructure/vqa/images/input/zh_val_0.jpg diff --git a/pplayout/images/input/zh_val_42.jpg b/ppstructure/vqa/images/input/zh_val_42.jpg similarity index 100% rename from pplayout/images/input/zh_val_42.jpg rename to ppstructure/vqa/images/input/zh_val_42.jpg diff --git a/pplayout/images/result_ser/zh_val_0_ser.jpg b/ppstructure/vqa/images/result_ser/zh_val_0_ser.jpg similarity index 100% rename from pplayout/images/result_ser/zh_val_0_ser.jpg rename to ppstructure/vqa/images/result_ser/zh_val_0_ser.jpg diff --git a/pplayout/images/result_ser/zh_val_42_ser.jpg b/ppstructure/vqa/images/result_ser/zh_val_42_ser.jpg similarity index 100% rename from pplayout/images/result_ser/zh_val_42_ser.jpg rename to ppstructure/vqa/images/result_ser/zh_val_42_ser.jpg diff --git a/pplayout/infer_ser.py b/ppstructure/vqa/infer_ser.py similarity index 100% rename from pplayout/infer_ser.py rename to ppstructure/vqa/infer_ser.py diff --git a/pplayout/infer_ser_e2e.py b/ppstructure/vqa/infer_ser_e2e.py similarity index 100% rename from pplayout/infer_ser_e2e.py rename to ppstructure/vqa/infer_ser_e2e.py diff --git a/pplayout/labels/labels_ser.txt b/ppstructure/vqa/labels/labels_ser.txt similarity index 100% rename from pplayout/labels/labels_ser.txt rename to ppstructure/vqa/labels/labels_ser.txt diff --git a/pplayout/requirements.txt b/ppstructure/vqa/requirements.txt similarity index 100% rename from pplayout/requirements.txt rename to ppstructure/vqa/requirements.txt diff --git a/pplayout/train_ser.py b/ppstructure/vqa/train_ser.py similarity index 100% rename from pplayout/train_ser.py rename to ppstructure/vqa/train_ser.py diff --git a/pplayout/utils.py b/ppstructure/vqa/utils.py similarity index 100% rename from pplayout/utils.py rename to ppstructure/vqa/utils.py diff --git a/pplayout/xfun.py b/ppstructure/vqa/xfun.py similarity index 100% rename from pplayout/xfun.py rename to ppstructure/vqa/xfun.py