Merge pull request #2347 from littletomatodonkey/dyg/fix_pre_rec

fix eval res vary for different times

PPOCRLabel/README.md

@@ -8,6 +8,10 @@ PPOCRLabel is a semi-automatic graphic annotation tool suitable for OCR field, w
 
 ### Recent Update
 
+- 2021.2.5: New batch processing and undo functions (by [Evezerest](https://github.com/Evezerest)):
+  - Batch processing function: Press and hold the Ctrl key to select the box, you can move, copy, and delete in batches.
+  - Undo function: In the process of drawing a four-point label box or after editing the box, press Ctrl+Z to undo the previous operation.
+  - Fix image rotation and size problems, optimize the process of editing the mark frame (by [ninetailskim](https://github.com/ninetailskim)、 [edencfc](https://github.com/edencfc)).
 - 2021.1.11: Optimize the labeling experience (by [edencfc](https://github.com/edencfc)),
   - Users can choose whether to pop up the label input dialog after drawing the detection box in "View - Pop-up Label Input Dialog".
   - The recognition result scrolls synchronously when users click related detection box.
@@ -16,7 +20,6 @@ PPOCRLabel is a semi-automatic graphic annotation tool suitable for OCR field, w
 
 ### TODO:
 - Lock box mode: For the same scene data, the size and position of the locked detection box can be transferred between different pictures.
-- Experience optimization: Add undo, batch operation include move, copy, delete and so on, optimize the annotation process.
 
 ## Installation
 
@@ -76,12 +79,11 @@ python3 PPOCRLabel.py
 
 7. Double click the result in 'recognition result' list to manually change inaccurate recognition results.
 
-8. Click "Check", the image status will switch to "√",then the program automatically jump to the next(The results will not be written directly to the file at this time).
+8. Click "Check", the image status will switch to "√",then the program automatically jump to the next.
 
 9. Click "Delete Image" and the image will be deleted to the recycle bin.
 
-10. Labeling result: the user can save manually through the menu "File - Save Label", while the program will also save automatically after every 5 images confirmed by the user.the manually checked label will be stored in *Label.txt* under the opened picture folder.
-    Click "PaddleOCR"-"Save Recognition Results" in the menu bar, the recognition training data of such pictures will be saved in the *crop_img* folder, and the recognition label will be saved in *rec_gt.txt*<sup>[4]</sup>.
+10. Labeling result: the user can save manually through the menu "File - Save Label", while the program will also save automatically if "File - Auto Save Label Mode" is selected. The manually checked label will be stored in *Label.txt* under the opened picture folder. Click "PaddleOCR"-"Save Recognition Results" in the menu bar, the recognition training data of such pictures will be saved in the *crop_img* folder, and the recognition label will be saved in *rec_gt.txt*<sup>[4]</sup>.
 
 ### Note
 
@@ -89,8 +91,7 @@ python3 PPOCRLabel.py
 
 [2] The image status indicates whether the user has saved the image manually. If it has not been saved manually it is "X", otherwise it is "√", PPOCRLabel will not relabel pictures with a status of "√".
 
-[3] After clicking "Re-recognize", the model will overwrite ALL recognition results in the picture.
-Therefore, if the recognition result has been manually changed before, it may change after re-recognition.
+[3] After clicking "Re-recognize", the model will overwrite ALL recognition results in the picture. Therefore, if the recognition result has been manually changed before, it may change after re-recognition.
 
 [4] The files produced by PPOCRLabel can be found under the opened picture folder including the following, please do not manually change the contents, otherwise it will cause the program to be abnormal.
 
@@ -106,22 +107,23 @@ Therefore, if the recognition result has been manually changed before, it may ch
 
 ### Shortcut keys
 
-| Shortcut keys    | Description                                      |
-| ---------------- | ------------------------------------------------ |
-| Ctrl + shift + A | Automatically label all unchecked images         |
-| Ctrl + shift + R | Re-recognize all the labels of the current image |
-| W                | Create a rect box                                |
-| Q                | Create a four-points box                         |
-| Ctrl + E         | Edit label of the selected box                   |
-| Ctrl + R         | Re-recognize the selected box                    |
-| Backspace        | Delete the selected box                          |
-| Ctrl + V         | Check image                                      |
-| Ctrl + Shift + d | Delete image                                     |
-| D                | Next image                                       |
-| A                | Previous image                                   |
-| Ctrl++           | Zoom in                                          |
-| Ctrl--           | Zoom out                                         |
-| ↑→↓←             | Move selected box                                |
+| Shortcut keys            | Description                                      |
+| ------------------------ | ------------------------------------------------ |
+| Ctrl + Shift + R         | Re-recognize all the labels of the current image |
+| W                        | Create a rect box                                |
+| Q                        | Create a four-points box                         |
+| Ctrl + E                 | Edit label of the selected box                   |
+| Ctrl + R                 | Re-recognize the selected box                    |
+| Ctrl + C                 | Copy and paste the selected box                  |
+| Ctrl + Left Mouse Button | Multi select the label box                       |
+| Backspace                | Delete the selected box                          |
+| Ctrl + V                 | Check image                                      |
+| Ctrl + Shift + d         | Delete image                                     |
+| D                        | Next image                                       |
+| A                        | Previous image                                   |
+| Ctrl++                   | Zoom in                                          |
+| Ctrl--                   | Zoom out                                         |
+| ↑→↓←                     | Move selected box                                |
 
 ### Built-in Model
 
@@ -136,7 +138,7 @@ Therefore, if the recognition result has been manually changed before, it may ch
 
 PPOCRLabel supports three ways to save Label.txt
 
-- Automatically save: When it detects that the user has manually checked 5 pictures, the program automatically writes the annotations into Label.txt. The user can change the value of ``self.autoSaveNum`` in ``PPOCRLabel.py`` to set the number of images to be automatically saved after confirmation.
+- Automatically save: After selecting "File - Auto Save Label Mode", the program will automatically write the annotations into Label.txt every time the user confirms an image. If this option is not turned on, it will be automatically saved after detecting that the user has manually checked 5 images.
 - Manual save: Click "File-Save Marking Results" to manually save the label.
 - Close application save
 
@@ -160,11 +162,11 @@ For some data that are difficult to recognize, the recognition results will not
     ```
     pyrcc5 -o libs/resources.py resources.qrc
     ```
-- If you get an error ``` module 'cv2' has no attribute 'INTER_NEAREST'```, you need to delete all opencv related packages first, and then reinstall the 4.2.0.32 version of headless opencv
+- If you get an error ``` module 'cv2' has no attribute 'INTER_NEAREST'```, you need to delete all opencv related packages first, and then reinstall the 4.2.0.32  version of headless opencv
     ```
     pip install opencv-contrib-python-headless==4.2.0.32
     ```
 
 ### Related
 
-1.[Tzutalin. LabelImg. Git code (2015)](https://github.com/tzutalin/labelImg)
+1.[Tzutalin. LabelImg. Git code (2015)](https://github.com/tzutalin/labelImg)

PPOCRLabel/README_ch.md

@@ -8,6 +8,10 @@ PPOCRLabel是一款适用于OCR领域的半自动化图形标注工具，内置P
 
 #### 近期更新
 
+- 2021.2.5：新增批处理与撤销功能（by [Evezerest](https://github.com/Evezerest))
+  - 批处理功能：按住Ctrl键选择标记框后可批量移动、复制、删除。
+  - 撤销功能：在绘制四点标注框过程中或对框进行编辑操作后，按下Ctrl+Z可撤销上一部操作。
+  - 修复图像旋转和尺寸问题、优化编辑标记框过程（by [ninetailskim](https://github.com/ninetailskim)、 [edencfc](https://github.com/edencfc)）
 - 2021.1.11：优化标注体验（by [edencfc](https://github.com/edencfc)）：
   - 用户可在“视图 - 弹出标记输入框”选择在画完检测框后标记输入框是否弹出。
   - 识别结果与检测框同步滚动。
@@ -17,9 +21,8 @@ PPOCRLabel是一款适用于OCR领域的半自动化图形标注工具，内置P
 #### 尽请期待
 
 - 锁定框模式：针对同一场景数据，被锁定的检测框的大小与位置能在不同图片之间传递。
-- 体验优化：增加撤销操作，批量移动、复制、删除等功能。优化标注流程。
 
-如果您对以上内容感兴趣或对完善工具有不一样的想法，欢迎加入我们的队伍与我们共同开发
+如果您对以上内容感兴趣或对完善工具有不一样的想法，欢迎加入我们的SIG队伍与我们共同开发。可以在[此处](https://github.com/PaddlePaddle/PaddleOCR/issues/1728)完成问卷和前置任务，经过我们确认相关内容后即可正式加入，享受SIG福利，共同为OCR开源事业贡献（特别说明：针对PPOCRLabel的改进也属于PaddleOCR前置任务）
 
 
 ## 安装
@@ -65,9 +68,9 @@ python3 PPOCRLabel.py --lang ch
 5. 标记框绘制完成后，用户点击 “确认”，检测框会先被预分配一个 “待识别” 标签。
 6. 重新识别：将图片中的所有检测画绘制/调整完成后，点击 “重新识别”，PPOCR模型会对当前图片中的**所有检测框**重新识别<sup>[3]</sup>。
 7. 内容更改：双击识别结果，对不准确的识别结果进行手动更改。
-8. 确认标记：点击 “确认”，图片状态切换为 “√”，跳转至下一张（此时不会直接将结果写入文件）。
+8. **确认标记**：点击 “确认”，图片状态切换为 “√”，跳转至下一张。
 9. 删除：点击 “删除图像”，图片将会被删除至回收站。
-10. 保存结果：用户可以通过菜单中“文件-保存标记结果”手动保存，同时程序也会在用户每确认5张图片后自动保存一次。手动确认过的标记将会被存放在所打开图片文件夹下的*Label.txt*中。在菜单栏点击 “文件” - "保存识别结果"后，会将此类图片的识别训练数据保存在*crop_img*文件夹下，识别标签保存在*rec_gt.txt*中<sup>[4]</sup>。
+10. 保存结果：用户可以通过菜单中“文件-保存标记结果”手动保存，同时也可以点击“文件 - 自动保存标记结果”开启自动保存。手动确认过的标记将会被存放在所打开图片文件夹下的*Label.txt*中。在菜单栏点击 “文件” - "保存识别结果"后，会将此类图片的识别训练数据保存在*crop_img*文件夹下，识别标签保存在*rec_gt.txt*中<sup>[4]</sup>。
 
 ### 注意
 
@@ -93,12 +96,13 @@ python3 PPOCRLabel.py --lang ch
 
 | 快捷键           | 说明                         |
 | ---------------- | ---------------------------- |
-| Ctrl + shift + A | 自动标注所有未确认过的图片   |
 | Ctrl + shift + R | 对当前图片的所有标记重新识别 |
 | W                | 新建矩形框                   |
 | Q                | 新建四点框                   |
 | Ctrl + E         | 编辑所选框标签               |
 | Ctrl + R         | 重新识别所选标记             |
+| Ctrl + C         | 复制并粘贴选中的标记框       |
+| Ctrl + 鼠标左键  | 多选标记框                   |
 | Backspace        | 删除所选框                   |
 | Ctrl + V         | 确认本张图片标记             |
 | Ctrl + Shift + d | 删除本张图片                 |
@@ -120,7 +124,7 @@ python3 PPOCRLabel.py --lang ch
 
 PPOCRLabel支持三种保存方式：
 
-- 程序自动保存：当检测到用户手动确认过5张图片后，程序自动将标记结果写入Label.txt中。其中用户可通过更改```PPOCRLabel.py```中的```self.autoSaveNum```的数值设置确认几张图片后进行自动保存。
+- 自动保存：点击“文件 - 自动保存标记结果”后，用户每确认过一张图片，程序自动将标记结果写入Label.txt中。若未开启此选项，则检测到用户手动确认过5张图片后进行自动保存。
 - 手动保存：点击“文件 - 保存标记结果”手动保存标记。
 - 关闭应用程序保存
 

PPOCRLabel/libs/shape.py

@@ -82,7 +82,7 @@ class Shape(object):
         return False
 
     def addPoint(self, point):
-        if not self.reachMaxPoints():
+        if not self.reachMaxPoints():  # 4个点时发出close信号
             self.points.append(point)
 
     def popPoint(self):

PPOCRLabel/resources/strings/strings-zh-CN.properties

@@ -96,4 +96,7 @@ hideBox=隐藏所有标注
 showBox=显示所有标注
 saveLabel=保存标记结果
 singleRe=重识别此区块
-labelDialogOption=弹出标记输入框
+labelDialogOption=弹出标记输入框
+undo=撤销
+undoLastPoint=撤销上个点
+autoSaveMode=自动保存标记结果

PPOCRLabel/resources/strings/strings.properties

@@ -96,4 +96,7 @@ hideBox=Hide All Box
 showBox=Show All Box
 saveLabel=Save Label
 singleRe=Re-recognition RectBox
-labelDialogOption=Pop-up Label Input Dialog
+labelDialogOption=Pop-up Label Input Dialog
+undo=Undo
+undoLastPoint=Undo Last Point
+autoSaveMode=Auto Save Label Mode

README.md

@@ -9,7 +9,6 @@ PaddleOCR supports both dynamic graph and static graph programming paradigm
 - Static graph: develop branch
 
 **Recent updates**
-- 2021.2.8 Release PaddleOCRv2.0(branch release/2.0) and set as default branch. Check release note here: https://github.com/PaddlePaddle/PaddleOCR/releases/tag/v2.0.0
 - 2021.1.21 update more than 25+ multilingual recognition models [models list](./doc/doc_en/models_list_en.md), including：English, Chinese, German, French, Japanese，Spanish，Portuguese Russia Arabic and so on.  Models for more languages will continue to be updated [Develop Plan](https://github.com/PaddlePaddle/PaddleOCR/issues/1048).
 - 2020.12.15 update Data synthesis tool, i.e., [Style-Text](./StyleText/README.md)，easy to synthesize a large number of images which are similar to the target scene image.
 - 2020.11.25 Update a new data annotation tool, i.e., [PPOCRLabel](./PPOCRLabel/README.md), which is helpful to improve the labeling efficiency. Moreover, the labeling results can be used in training of the PP-OCR system directly.
@@ -43,7 +42,7 @@ The above pictures are the visualizations of the general ppocr_server model. For
 - Scan the QR code below with your Wechat, you can access to official technical exchange group. Look forward to your participation.
 
 <div align="center">
-<img src="./doc/joinus.PNG"  width = "200" height = "200" />
+<img src="https://raw.githubusercontent.com/PaddlePaddle/PaddleOCR/release/2.0/doc/joinus.PNG"  width = "200" height = "200" />
 </div>
 
 
@@ -94,7 +93,7 @@ For a new language request, please refer to [Guideline for new language_requests
     - [Quick Inference Based on PIP](./doc/doc_en/whl_en.md)
     - [Python Inference](./doc/doc_en/inference_en.md)
     - [C++ Inference](./deploy/cpp_infer/readme_en.md)
-    - [Serving](./deploy/hubserving/readme_en.md)
+    - [Serving](./deploy/pdserving/README.md)
     - [Mobile](https://github.com/PaddlePaddle/PaddleOCR/blob/develop/deploy/lite/readme_en.md)
     - [Benchmark](./doc/doc_en/benchmark_en.md)  
 - Data Annotation and Synthesis

README_ch.md

@@ -8,8 +8,7 @@ PaddleOCR同时支持动态图与静态图两种编程范式
 - 静态图版本：develop分支
 
 **近期更新**
-- 2021.3.22 [FAQ](./doc/doc_ch/FAQ.md)新增5个高频问题，总数193个，每周一都会更新，欢迎大家持续关注。
-- 2021.2.8 正式发布PaddleOCRv2.0(branch release/2.0)并设置为推荐用户使用的默认分支. 发布的详细内容，请参考: https://github.com/PaddlePaddle/PaddleOCR/releases/tag/v2.0.0
+- 2021.2.1 [FAQ](./doc/doc_ch/FAQ.md)新增5个高频问题，总数162个，每周一都会更新，欢迎大家持续关注。
 - 2021.1.26,28,29 PaddleOCR官方研发团队带来技术深入解读三日直播课，1月26日、28日、29日晚上19:30，[直播地址](https://live.bilibili.com/21689802)
 - 2021.1.21 更新多语言识别模型，目前支持语种超过27种，[多语言模型下载](./doc/doc_ch/models_list.md)，包括中文简体、中文繁体、英文、法文、德文、韩文、日文、意大利文、西班牙文、葡萄牙文、俄罗斯文、阿拉伯文等，后续计划可以参考[多语言研发计划](https://github.com/PaddlePaddle/PaddleOCR/issues/1048)
 - 2020.12.15 更新数据合成工具[Style-Text](./StyleText/README_ch.md)，可以批量合成大量与目标场景类似的图像，在多个场景验证，效果明显提升。
@@ -47,7 +46,7 @@ PaddleOCR同时支持动态图与静态图两种编程范式
 - 微信扫描二维码加入官方交流群，获得更高效的问题答疑，与各行各业开发者充分交流，期待您的加入。
 
 <div align="center">
-<img src="./doc/joinus.PNG"  width = "200" height = "200" />
+<img src="https://raw.githubusercontent.com/PaddlePaddle/PaddleOCR/release/2.0/doc/joinus.PNG"  width = "200" height = "200" />
 </div>
 
 ## 快速体验
@@ -89,7 +88,7 @@ PaddleOCR同时支持动态图与静态图两种编程范式
     - [基于pip安装whl包快速推理](./doc/doc_ch/whl.md)
     - [基于Python脚本预测引擎推理](./doc/doc_ch/inference.md)
     - [基于C++预测引擎推理](./deploy/cpp_infer/readme.md)
-    - [服务化部署](./deploy/hubserving/readme.md)
+    - [服务化部署](./deploy/pdserving/README_CN.md)
     - [端侧部署](https://github.com/PaddlePaddle/PaddleOCR/blob/develop/deploy/lite/readme.md)
     - [Benchmark](./doc/doc_ch/benchmark.md)
 - 数据集
@@ -104,8 +103,8 @@ PaddleOCR同时支持动态图与静态图两种编程范式
 - [效果展示](#效果展示)
 - FAQ
     - [【精选】OCR精选10个问题](./doc/doc_ch/FAQ.md)
-    - [【理论篇】OCR通用37个问题](./doc/doc_ch/FAQ.md)
-    - [【实战篇】PaddleOCR实战141个问题](./doc/doc_ch/FAQ.md)
+    - [【理论篇】OCR通用32个问题](./doc/doc_ch/FAQ.md)
+    - [【实战篇】PaddleOCR实战110个问题](./doc/doc_ch/FAQ.md)
 - [技术交流群](#欢迎加入PaddleOCR技术交流群)
 - [参考文献](./doc/doc_ch/reference.md)
 - [许可证书](#许可证书)

StyleText/README.md

@@ -98,7 +98,7 @@ What's more, the medium result `fake_bg.jpg` will also be saved, which is the ba
 </div>
 
 
-`fake_text.jpg` is the generated image with the same font style as `Style Input`.
+`fake_text.jpg` * `fake_text.jpg` is the generated image with the same font style as `Style Input`.
 
 
 <div align="center">

StyleText/engine/predictors.py

@@ -38,7 +38,15 @@ class StyleTextRecPredictor(object):
         self.std = config["Predictor"]["std"]
         self.expand_result = config["Predictor"]["expand_result"]
 
-    def predict(self, style_input, text_input):
+    def reshape_to_same_height(self, img_list):
+        h = img_list[0].shape[0]
+        for idx in range(1, len(img_list)):
+            new_w = round(1.0 * img_list[idx].shape[1] /
+                          img_list[idx].shape[0] * h)
+            img_list[idx] = cv2.resize(img_list[idx], (new_w, h))
+        return img_list
+
+    def predict_single_image(self, style_input, text_input):
         style_input = self.rep_style_input(style_input, text_input)
         tensor_style_input = self.preprocess(style_input)
         tensor_text_input = self.preprocess(text_input)
@@ -64,6 +72,21 @@ class StyleTextRecPredictor(object):
             "fake_bg": fake_bg,
         }
 
+    def predict(self, style_input, text_input_list):
+        if not isinstance(text_input_list, (tuple, list)):
+            return self.predict_single_image(style_input, text_input_list)
+
+        synth_result_list = []
+        for text_input in text_input_list:
+            synth_result = self.predict_single_image(style_input, text_input)
+            synth_result_list.append(synth_result)
+
+        for key in synth_result:
+            res = [r[key] for r in synth_result_list]
+            res = self.reshape_to_same_height(res)
+            synth_result[key] = np.concatenate(res, axis=1)
+        return synth_result
+
     def preprocess(self, img):
         img = (img.astype('float32') * self.scale - self.mean) / self.std
         img_height, img_width, channel = img.shape

StyleText/engine/synthesisers.py

@@ -12,6 +12,8 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 import os
+import numpy as np
+import cv2
 
 from utils.config import ArgsParser, load_config, override_config
 from utils.logging import get_logger
@@ -36,8 +38,9 @@ class ImageSynthesiser(object):
         self.predictor = getattr(predictors, predictor_method)(self.config)
 
     def synth_image(self, corpus, style_input, language="en"):
-        corpus, text_input = self.text_drawer.draw_text(corpus, language)
-        synth_result = self.predictor.predict(style_input, text_input)
+        corpus_list, text_input_list = self.text_drawer.draw_text(
+            corpus, language, style_input_width=style_input.shape[1])
+        synth_result = self.predictor.predict(style_input, text_input_list)
         return synth_result
 
 
@@ -59,12 +62,15 @@ class DatasetSynthesiser(ImageSynthesiser):
         for i in range(self.output_num):
             style_data = self.style_sampler.sample()
             style_input = style_data["image"]
-            corpus_language, text_input_label = self.corpus_generator.generate(
-            )
-            text_input_label, text_input = self.text_drawer.draw_text(
-                text_input_label, corpus_language)
+            corpus_language, text_input_label = self.corpus_generator.generate()
+            text_input_label_list, text_input_list = self.text_drawer.draw_text(
+                text_input_label,
+                corpus_language,
+                style_input_width=style_input.shape[1])
 
-            synth_result = self.predictor.predict(style_input, text_input)
+            text_input_label = "".join(text_input_label_list)
+
+            synth_result = self.predictor.predict(style_input, text_input_list)
             fake_fusion = synth_result["fake_fusion"]
             self.writer.save_image(fake_fusion, text_input_label)
         self.writer.save_label()

StyleText/engine/text_drawers.py

@@ -1,5 +1,6 @@
 from PIL import Image, ImageDraw, ImageFont
 import numpy as np
+import cv2
 from utils.logging import get_logger
 
 
@@ -28,7 +29,11 @@ class StdTextDrawer(object):
         else:
             return int((self.height - 4)**2 / font_height)
 
-    def draw_text(self, corpus, language="en", crop=True):
+    def draw_text(self,
+                  corpus,
+                  language="en",
+                  crop=True,
+                  style_input_width=None):
         if language not in self.support_languages:
             self.logger.warning(
                 "language {} not supported, use en instead.".format(language))
@@ -37,21 +42,43 @@ class StdTextDrawer(object):
             width = min(self.max_width, len(corpus) * self.height) + 4
         else:
             width = len(corpus) * self.height + 4
-        bg = Image.new("RGB", (width, self.height), color=(127, 127, 127))
-        draw = ImageDraw.Draw(bg)
-
-        char_x = 2
-        font = self.font_dict[language]
-        for i, char_i in enumerate(corpus):
-            char_size = font.getsize(char_i)[0]
-            draw.text((char_x, 2), char_i, fill=(0, 0, 0), font=font)
-            char_x += char_size
-            if char_x >= width:
-                corpus = corpus[0:i + 1]
-                self.logger.warning("corpus length exceed limit: {}".format(
-                    corpus))
+
+        if style_input_width is not None:
+            width = min(width, style_input_width)
+
+        corpus_list = []
+        text_input_list = []
+
+        while len(corpus) != 0:
+            bg = Image.new("RGB", (width, self.height), color=(127, 127, 127))
+            draw = ImageDraw.Draw(bg)
+            char_x = 2
+            font = self.font_dict[language]
+            i = 0
+            while i < len(corpus):
+                char_i = corpus[i]
+                char_size = font.getsize(char_i)[0]
+                # split when char_x exceeds char size and index is not 0 (at least 1 char should be wroten on the image)
+                if char_x + char_size >= width and i != 0:
+                    text_input = np.array(bg).astype(np.uint8)
+                    text_input = text_input[:, 0:char_x, :]
+
+                    corpus_list.append(corpus[0:i])
+                    text_input_list.append(text_input)
+                    corpus = corpus[i:]
+                    break
+                draw.text((char_x, 2), char_i, fill=(0, 0, 0), font=font)
+                char_x += char_size
+
+                i += 1
+            # the whole text is shorter than style input
+            if i == len(corpus):
+                text_input = np.array(bg).astype(np.uint8)
+                text_input = text_input[:, 0:char_x, :]
+
+                corpus_list.append(corpus[0:i])
+                text_input_list.append(text_input)
+                corpus = corpus[i:]
                 break
 
-        text_input = np.array(bg).astype(np.uint8)
-        text_input = text_input[:, 0:char_x, :]
-        return corpus, text_input
+        return corpus_list, text_input_list

StyleText/examples/image_list.txt

@@ -1,2 +1,2 @@
 style_images/1.jpg	NEATNESS
-style_images/2.jpg	卡丹鑫宇通
+style_images/2.jpg	锁店君和宾馆

configs/cls/cls_mv3.yml

@@ -7,7 +7,6 @@ Global:
   save_epoch_step: 3
   # evaluation is run every 5000 iterations after the 4000th iteration
   eval_batch_step: [0, 1000]
-  # if pretrained_model is saved in static mode, load_static_weights must set to True
   cal_metric_during_train: True
   pretrained_model:
   checkpoints:
@@ -93,4 +92,4 @@ Eval:
     shuffle: False
     drop_last: False
     batch_size_per_card: 512
-    num_workers: 4
+    num_workers: 4

configs/det/ch_ppocr_v2.0/ch_det_mv3_db_v2.0.yml

@@ -7,7 +7,10 @@ Global:
   save_epoch_step: 1200
   # evaluation is run every 5000 iterations after the 4000th iteration
   eval_batch_step: [3000, 2000]
-  # if pretrained_model is saved in static mode, load_static_weights must set to True
+  # 1. If pretrained_model is saved in static mode, such as classification pretrained model
+  #    from static branch, load_static_weights must be set as True.
+  # 2. If you want to finetune the pretrained models we provide in the docs,
+  #    you should set load_static_weights as False.
   load_static_weights: True
   cal_metric_during_train: False
   pretrained_model: ./pretrain_models/MobileNetV3_large_x0_5_pretrained

configs/det/ch_ppocr_v2.0/ch_det_res18_db_v2.0.yml

@@ -7,7 +7,10 @@ Global:
   save_epoch_step: 1200
   # evaluation is run every 5000 iterations after the 4000th iteration
   eval_batch_step: [3000, 2000]
-  # if pretrained_model is saved in static mode, load_static_weights must set to True
+  # 1. If pretrained_model is saved in static mode, such as classification pretrained model
+  #    from static branch, load_static_weights must be set as True.
+  # 2. If you want to finetune the pretrained models we provide in the docs,
+  #    you should set load_static_weights as False.
   load_static_weights: True
   cal_metric_during_train: False
   pretrained_model: ./pretrain_models/ResNet18_vd_pretrained

configs/det/det_mv3_db.yml

@@ -7,7 +7,10 @@ Global:
   save_epoch_step: 1200
   # evaluation is run every 2000 iterations
   eval_batch_step: [0, 2000]
-  # if pretrained_model is saved in static mode, load_static_weights must set to True
+  # 1. If pretrained_model is saved in static mode, such as classification pretrained model
+  #    from static branch, load_static_weights must be set as True.
+  # 2. If you want to finetune the pretrained models we provide in the docs,
+  #    you should set load_static_weights as False.
   load_static_weights: True
   cal_metric_during_train: False
   pretrained_model: ./pretrain_models/MobileNetV3_large_x0_5_pretrained

configs/det/det_mv3_east.yml

@@ -7,7 +7,10 @@ Global:
   save_epoch_step: 1000
   # evaluation is run every 5000 iterations after the 4000th iteration
   eval_batch_step: [4000, 5000]
-  # if pretrained_model is saved in static mode, load_static_weights must set to True
+  # 1. If pretrained_model is saved in static mode, such as classification pretrained model
+  #    from static branch, load_static_weights must be set as True.
+  # 2. If you want to finetune the pretrained models we provide in the docs,
+  #    you should set load_static_weights as False.
   load_static_weights: True
   cal_metric_during_train: False
   pretrained_model: ./pretrain_models/MobileNetV3_large_x0_5_pretrained

configs/det/det_r50_vd_db.yml

@@ -7,7 +7,10 @@ Global:
   save_epoch_step: 1200
   # evaluation is run every 2000 iterations
   eval_batch_step: [0,2000]
-  # if pretrained_model is saved in static mode, load_static_weights must set to True
+  # 1. If pretrained_model is saved in static mode, such as classification pretrained model
+  #    from static branch, load_static_weights must be set as True.
+  # 2. If you want to finetune the pretrained models we provide in the docs,
+  #    you should set load_static_weights as False.
   load_static_weights: True
   cal_metric_during_train: False
   pretrained_model: ./pretrain_models/ResNet50_vd_ssld_pretrained

configs/det/det_r50_vd_east.yml

@@ -7,7 +7,10 @@ Global:
   save_epoch_step: 1000
   # evaluation is run every 5000 iterations after the 4000th iteration
   eval_batch_step: [4000, 5000]
-  # if pretrained_model is saved in static mode, load_static_weights must set to True
+  # 1. If pretrained_model is saved in static mode, such as classification pretrained model
+  #    from static branch, load_static_weights must be set as True.
+  # 2. If you want to finetune the pretrained models we provide in the docs,
+  #    you should set load_static_weights as False.
   load_static_weights: True
   cal_metric_during_train: False
   pretrained_model: ./pretrain_models/ResNet50_vd_pretrained/

configs/det/det_r50_vd_sast_icdar15.yml

@@ -7,7 +7,10 @@ Global:
   save_epoch_step: 1000
   # evaluation is run every 5000 iterations after the 4000th iteration
   eval_batch_step: [4000, 5000]
-  # if pretrained_model is saved in static mode, load_static_weights must set to True
+  # 1. If pretrained_model is saved in static mode, such as classification pretrained model
+  #    from static branch, load_static_weights must be set as True.
+  # 2. If you want to finetune the pretrained models we provide in the docs,
+  #    you should set load_static_weights as False.
   load_static_weights: True
   cal_metric_during_train: False
   pretrained_model: ./pretrain_models/ResNet50_vd_ssld_pretrained/

configs/det/det_r50_vd_sast_totaltext.yml

@@ -7,7 +7,10 @@ Global:
   save_epoch_step: 1000
   # evaluation is run every 5000 iterations after the 4000th iteration
   eval_batch_step: [4000, 5000]
-  # if pretrained_model is saved in static mode, load_static_weights must set to True
+  # 1. If pretrained_model is saved in static mode, such as classification pretrained model
+  #    from static branch, load_static_weights must be set as True.
+  # 2. If you want to finetune the pretrained models we provide in the docs,
+  #    you should set load_static_weights as False.
   load_static_weights: True
   cal_metric_during_train: False
   pretrained_model: ./pretrain_models/ResNet50_vd_ssld_pretrained/

configs/rec/ch_ppocr_v2.0/rec_chinese_common_train_v2.0.yml

@@ -7,7 +7,6 @@ Global:
   save_epoch_step: 3
   # evaluation is run every 5000 iterations after the 4000th iteration
   eval_batch_step: [0, 2000]
-  # if pretrained_model is saved in static mode, load_static_weights must set to True
   cal_metric_during_train: True
   pretrained_model:
   checkpoints: