fixed conflicts 20220801

PPOCRLabel/PPOCRLabel.py

@@ -28,7 +28,7 @@ from PyQt5.QtCore import QSize, Qt, QPoint, QByteArray, QTimer, QFileInfo, QPoin
 from PyQt5.QtGui import QImage, QCursor, QPixmap, QImageReader
 from PyQt5.QtWidgets import QMainWindow, QListWidget, QVBoxLayout, QToolButton, QHBoxLayout, QDockWidget, QWidget, \
     QSlider, QGraphicsOpacityEffect, QMessageBox, QListView, QScrollArea, QWidgetAction, QApplication, QLabel, QGridLayout, \
-    QFileDialog, QListWidgetItem, QComboBox, QDialog, QAbstractItemView
+    QFileDialog, QListWidgetItem, QComboBox, QDialog, QAbstractItemView, QSizePolicy
 
 __dir__ = os.path.dirname(os.path.abspath(__file__))
 
@@ -227,6 +227,21 @@ class MainWindow(QMainWindow):
         listLayout.addWidget(leftTopToolBoxContainer)
 
         #  ================== Label List  ==================
+        labelIndexListlBox = QHBoxLayout()
+
+        # Create and add a widget for showing current label item index
+        self.indexList = QListWidget()
+        self.indexList.setMaximumSize(30, 16777215) # limit max width
+        self.indexList.setEditTriggers(QAbstractItemView.NoEditTriggers) # no editable
+        self.indexList.itemSelectionChanged.connect(self.indexSelectionChanged)
+        self.indexList.setVerticalScrollBarPolicy(Qt.ScrollBarAlwaysOff) # no scroll Bar
+        self.indexListDock = QDockWidget('No.', self)
+        self.indexListDock.setWidget(self.indexList)
+        self.indexListDock.setFeatures(QDockWidget.NoDockWidgetFeatures)
+        labelIndexListlBox.addWidget(self.indexListDock, 1)
+        # no margin between two boxes
+        labelIndexListlBox.setSpacing(0)
+        
         # Create and add a widget for showing current label items
         self.labelList = EditInList()
         labelListContainer = QWidget()
@@ -240,8 +255,8 @@ class MainWindow(QMainWindow):
         self.labelListDock = QDockWidget(self.labelListDockName, self)
         self.labelListDock.setWidget(self.labelList)
         self.labelListDock.setFeatures(QDockWidget.NoDockWidgetFeatures)
-        listLayout.addWidget(self.labelListDock)
-
+        labelIndexListlBox.addWidget(self.labelListDock, 10) # label list is wider than index list
+        
         # enable labelList drag_drop to adjust bbox order
         # 设置选择模式为单选  
         self.labelList.setSelectionMode(QAbstractItemView.SingleSelection)
@@ -256,6 +271,17 @@ class MainWindow(QMainWindow):
         # 触发放置
         self.labelList.model().rowsMoved.connect(self.drag_drop_happened)
 
+        labelIndexListContainer = QWidget()
+        labelIndexListContainer.setLayout(labelIndexListlBox)
+        listLayout.addWidget(labelIndexListContainer)
+
+        # labelList indexList同步滚动
+        self.labelListBar = self.labelList.verticalScrollBar()
+        self.indexListBar = self.indexList.verticalScrollBar()
+
+        self.labelListBar.valueChanged.connect(self.move_scrollbar)
+        self.indexListBar.valueChanged.connect(self.move_scrollbar)
+
         #  ================== Detection Box  ==================
         self.BoxList = QListWidget()
 
@@ -766,6 +792,7 @@ class MainWindow(QMainWindow):
         self.shapesToItemsbox.clear()
         self.labelList.clear()
         self.BoxList.clear()
+        self.indexList.clear()
         self.filePath = None
         self.imageData = None
         self.labelFile = None
@@ -1027,13 +1054,19 @@ class MainWindow(QMainWindow):
         for shape in self.canvas.selectedShapes:
             shape.selected = False
         self.labelList.clearSelection()
+        self.indexList.clearSelection()
         self.canvas.selectedShapes = selected_shapes
         for shape in self.canvas.selectedShapes:
             shape.selected = True
             self.shapesToItems[shape].setSelected(True)
             self.shapesToItemsbox[shape].setSelected(True)
+            index = self.labelList.indexFromItem(self.shapesToItems[shape]).row()
+            self.indexList.item(index).setSelected(True)
 
         self.labelList.scrollToItem(self.currentItem())  # QAbstractItemView.EnsureVisible
+        # map current label item to index item and select it
+        index = self.labelList.indexFromItem(self.currentItem()).row()
+        self.indexList.scrollToItem(self.indexList.item(index)) 
         self.BoxList.scrollToItem(self.currentBox())
 
         if self.kie_mode:
@@ -1066,12 +1099,18 @@ class MainWindow(QMainWindow):
         shape.paintIdx = self.displayIndexOption.isChecked()
 
         item = HashableQListWidgetItem(shape.label)
-        item.setFlags(item.flags() | Qt.ItemIsUserCheckable)
-        item.setCheckState(Qt.Unchecked) if shape.difficult else item.setCheckState(Qt.Checked)
+        # current difficult checkbox is disenble
+        # item.setFlags(item.flags() | Qt.ItemIsUserCheckable)
+        # item.setCheckState(Qt.Unchecked) if shape.difficult else item.setCheckState(Qt.Checked)
+
         # Checked means difficult is False
         # item.setBackground(generateColorByText(shape.label))
         self.itemsToShapes[item] = shape
         self.shapesToItems[shape] = item
+        # add current label item index before label string
+        current_index = QListWidgetItem(str(self.labelList.count()))
+        current_index.setTextAlignment(Qt.AlignHCenter)
+        self.indexList.addItem(current_index)
         self.labelList.addItem(item)
         # print('item in add label is ',[(p.x(), p.y()) for p in shape.points], shape.label)
 
@@ -1105,6 +1144,7 @@ class MainWindow(QMainWindow):
             del self.shapesToItemsbox[shape]
             del self.itemsToShapesbox[item]
             self.updateComboBox()
+        self.updateIndexList()
 
     def loadLabels(self, shapes):
         s = []
@@ -1156,6 +1196,13 @@ class MainWindow(QMainWindow):
 
         # self.comboBox.update_items(uniqueTextList)
 
+    def updateIndexList(self):
+        self.indexList.clear()
+        for i in range(self.labelList.count()):
+            string = QListWidgetItem(str(i))
+            string.setTextAlignment(Qt.AlignHCenter)
+            self.indexList.addItem(string)
+
     def saveLabels(self, annotationFilePath, mode='Auto'):
         # Mode is Auto means that labels will be loaded from self.result_dic totally, which is the output of ocr model
         annotationFilePath = ustr(annotationFilePath)
@@ -1211,6 +1258,10 @@ class MainWindow(QMainWindow):
         # fix copy and delete
         # self.shapeSelectionChanged(True)
 
+    def move_scrollbar(self, value):
+        self.labelListBar.setValue(value)
+        self.indexListBar.setValue(value)
+
     def labelSelectionChanged(self):
         if self._noSelectionSlot:
             return
@@ -1223,6 +1274,21 @@ class MainWindow(QMainWindow):
             else:
                 self.canvas.deSelectShape()
 
+    def indexSelectionChanged(self):
+        if self._noSelectionSlot:
+            return
+        if self.canvas.editing():
+            selected_shapes = []
+            for item in self.indexList.selectedItems():
+                # map index item to label item
+                index = self.indexList.indexFromItem(item).row()
+                item = self.labelList.item(index)
+                selected_shapes.append(self.itemsToShapes[item])
+            if selected_shapes:
+                self.canvas.selectShapes(selected_shapes)
+            else:
+                self.canvas.deSelectShape()
+
     def boxSelectionChanged(self):
         if self._noSelectionSlot:
             # self.BoxList.scrollToItem(self.currentBox(), QAbstractItemView.PositionAtCenter)
@@ -1517,6 +1583,7 @@ class MainWindow(QMainWindow):
             if self.labelList.count():
                 self.labelList.setCurrentItem(self.labelList.item(self.labelList.count() - 1))
                 self.labelList.item(self.labelList.count() - 1).setSelected(True)
+                self.indexList.item(self.labelList.count() - 1).setSelected(True)
 
             # show file list image count
             select_indexes = self.fileListWidget.selectedIndexes()
@@ -2015,12 +2082,14 @@ class MainWindow(QMainWindow):
         for shape in self.canvas.shapes:
             shape.paintLabel = self.displayLabelOption.isChecked()
             shape.paintIdx = self.displayIndexOption.isChecked()
+        self.canvas.repaint()
 
     def togglePaintIndexOption(self):
         self.displayLabelOption.setChecked(False)
         for shape in self.canvas.shapes:
             shape.paintLabel = self.displayLabelOption.isChecked()
             shape.paintIdx = self.displayIndexOption.isChecked()
+        self.canvas.repaint()
 
     def toogleDrawSquare(self):
         self.canvas.setDrawingShapeToSquare(self.drawSquaresOption.isChecked())
@@ -2254,6 +2323,7 @@ class MainWindow(QMainWindow):
                 self.itemsToShapesbox.clear()  # ADD
                 self.shapesToItemsbox.clear()
                 self.labelList.clear()
+                self.indexList.clear()
                 self.BoxList.clear()
                 self.result_dic = []
                 self.result_dic_locked = []
@@ -2665,6 +2735,7 @@ class MainWindow(QMainWindow):
     def undoShapeEdit(self):
         self.canvas.restoreShape()
         self.labelList.clear()
+        self.indexList.clear()
         self.BoxList.clear()
         self.loadShapes(self.canvas.shapes)
         self.actions.undo.setEnabled(self.canvas.isShapeRestorable)
@@ -2674,6 +2745,7 @@ class MainWindow(QMainWindow):
         for shape in shapes:
             self.addLabel(shape)
         self.labelList.clearSelection()
+        self.indexList.clearSelection()
         self._noSelectionSlot = False
         self.canvas.loadShapes(shapes, replace=replace)
         print("loadShapes")  # 1

PPOCRLabel/README.md

@@ -2,7 +2,7 @@ English | [简体中文](README_ch.md)
 
 # PPOCRLabel
 
-PPOCRLabel is a semi-automatic graphic annotation tool suitable for OCR field, with built-in PPOCR model to automatically detect and re-recognize data. It is written in python3 and pyqt5, supporting rectangular box, table and multi-point annotation modes. Annotations can be directly used for the training of PPOCR detection and recognition models.
+PPOCRLabel is a semi-automatic graphic annotation tool suitable for OCR field, with built-in PP-OCR model to automatically detect and re-recognize data. It is written in python3 and pyqt5, supporting rectangular box, table and multi-point annotation modes. Annotations can be directly used for the training of PP-OCR detection and recognition models.
 
 <img src="./data/gif/steps_en.gif" width="100%"/>
 
@@ -142,14 +142,18 @@ In PPOCRLabel, complete the text information labeling (text and position), compl
 labeling in the Excel file, the recommended steps are:
 
 1. Table annotation: After opening the table picture, click on the `Table Recognition` button in the upper right corner of PPOCRLabel, which will call the table recognition model in PP-Structure to automatically label 
-the table and pop up Excel at the same time.
-   
+    the table and pop up Excel at the same time.
+
 2. Change the recognition result: **label each cell** (i.e. the text in a cell is marked as a box). Right click on the box and click on `Cell Re-recognition`. 
    You can use the model to automatically recognise the text within a cell.
-   
+
 3. Mark the table structure: for each cell contains the text, **mark as any identifier (such as `1`) in Excel**, to ensure that the merged cell structure is same as the original picture.
 
-4. Export JSON format annotation: close all Excel files corresponding to table images, click `File`-`Export table JSON annotation` to obtain JSON annotation results.
+	> Note: If there are blank cells in the table, you also need to mark them with a bounding box so that the total number of cells is the same as in the image.
+
+4. ***Adjust cell order:*** Click on the menu  `View` - `Show Box Number` to show the box ordinal numbers, and drag all the results under the 'Recognition Results' column on the right side of the software interface to make the box numbers are arranged from left to right, top to bottom
+
+5. Export JSON format annotation: close all Excel files corresponding to table images, click `File`-`Export table JSON annotation` to obtain JSON annotation results.
 
 ### 2.3 Note
 
@@ -219,14 +223,7 @@ PPOCRLabel supports three ways to export Label.txt
 
 - Close application export
 
-
-### 3.4 Export Partial Recognition Results
-
-For some data that are difficult to recognize, the recognition results will not be exported by **unchecking** the corresponding tags in the recognition results checkbox. The unchecked recognition result is saved as `True` in the `difficult` variable in the label file `label.txt`.
-
-> *Note: The status of the checkboxes in the recognition results still needs to be saved manually by clicking Save Button.*
-
-### 3.5 Dataset division
+### 3.4 Dataset division
 
 - Enter the following command in the terminal to execute the dataset division script:
 
@@ -255,7 +252,7 @@ For some data that are difficult to recognize, the recognition results will not
     | ...
   ```
   
-### 3.6 Error message
+### 3.5 Error message
 
 - If paddleocr is installed with whl, it has a higher priority than calling PaddleOCR class with paddleocr.py, which may cause an exception if whl package is not updated.
 

PPOCRLabel/README_ch.md

@@ -7,8 +7,8 @@ PPOCRLabel是一款适用于OCR领域的半自动化图形标注工具，内置P
 <img src="./data/gif/steps.gif" width="100%"/>
 
 #### 近期更新
-- 2022.05：新增表格标注，使用方法见下方`2.2 表格标注`（by [whjdark](https://github.com/peterh0323); [Evezerest](https://github.com/Evezerest))
-- 2022.02：新增关键信息标注、优化标注体验（by [PeterH0323](https://github.com/peterh0323) ）
+- 2022.05：**新增表格标注**，使用方法见下方`2.2 表格标注`（by [whjdark](https://github.com/peterh0323); [Evezerest](https://github.com/Evezerest))
+- 2022.02：**新增关键信息标注**、优化标注体验（by [PeterH0323](https://github.com/peterh0323) ）
   - 新增：使用 `--kie` 进入 KIE 功能，用于打【检测+识别+关键字提取】的标签
   - 提升用户体验：新增文件与标记数目提示、优化交互、修复gpu使用等问题。
   - 新增功能：使用 `C` 和 `X` 对标记框进行旋转。
@@ -113,23 +113,29 @@ pip3 install dist/PPOCRLabel-1.0.2-py2.py3-none-any.whl -i https://mirror.baidu.
 
 1. 安装与运行：使用上述命令安装与运行程序。
 2. 打开文件夹：在菜单栏点击 “文件” - "打开目录" 选择待标记图片的文件夹<sup>[1]</sup>.
-3. 自动标注：点击 ”自动标注“，使用PPOCR超轻量模型对图片文件名前图片状态<sup>[2]</sup>为 “X” 的图片进行自动标注。
+3. 自动标注：点击 ”自动标注“，使用PP-OCR超轻量模型对图片文件名前图片状态<sup>[2]</sup>为 “X” 的图片进行自动标注。
 4. 手动标注：点击 “矩形标注”（推荐直接在英文模式下点击键盘中的 “W”)，用户可对当前图片中模型未检出的部分进行手动绘制标记框。点击键盘Q，则使用四点标注模式（或点击“编辑” - “四点标注”），用户依次点击4个点后，双击左键表示标注完成。
 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>。
+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”，PP-OCR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：单击识别结果，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。**
 9. 删除：点击 “删除图像”，图片将会被删除至回收站。
 10. 导出结果：用户可以通过菜单中“文件-导出标记结果”手动导出，同时也可以点击“文件 - 自动导出标记结果”开启自动导出。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” - "导出识别结果"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>。
 
 ### 2.2 表格标注
-表格标注针对表格的结构化提取，将图片中的表格转换为Excel格式，因此标注时需要配合外部软件打开Excel同时完成。
-在PPOCRLabel软件中完成表格中的文字信息标注（文字与位置）、在Excel文件中完成表格结构信息标注，推荐的步骤为：
+表格标注针对表格的结构化提取，将图片中的表格转换为Excel格式，因此标注时需要配合外部软件打开Excel同时完成。在PPOCRLabel软件中完成表格中的文字信息标注（文字与位置）、在Excel文件中完成表格结构信息标注，推荐的步骤为：
 1. 表格识别：打开表格图片后，点击软件右上角 `表格识别` 按钮，软件调用PP-Structure中的表格识别模型，自动为表格打标签，同时弹出Excel
-2. 更改识别结果：**以表格中的单元格为单位增加标注框**（即一个单元格内的文字都标记为一个框）。标注框上鼠标右键后点击 `单元格重识别` 
+
+2. 更改标注结果：**以表格中的单元格为单位增加标注框**（即一个单元格内的文字都标记为一个框）。标注框上鼠标右键后点击 `单元格重识别` 
    可利用模型自动识别单元格内的文字。
-3. 标注表格结构：将表格图像中有文字的单元格，**在Excel中标记为任意标识符（如`1`）**，保证Excel中的单元格合并情况与原图相同即可。
-4. 导出JSON格式：关闭所有表格图像对应的Excel，点击 `文件`-`导出表格JSON标注` 获得JSON标注结果。
+   
+   > 注意：如果表格中存在空白单元格，同样需要使用一个标注框将其标出，使得单元格总数与图像中保持一致。
+
+3. **调整单元格顺序：**点击软件`视图-显示框编号` 打开标注框序号，在软件界面右侧拖动 `识别结果` 一栏下的所有结果，使得标注框编号按照从左到右，从上到下的顺序排列
+   
+4. 标注表格结构：**在外部Excel软件中，将存在文字的单元格标记为任意标识符（如 `1` ）**，保证Excel中的单元格合并情况与原图相同即可（即不需要Excel中的单元格文字与图片中的文字完全相同）
+
+5. 导出JSON格式：关闭所有表格图像对应的Excel，点击 `文件`-`导出表格JSON标注` 获得JSON标注结果。
 
 ### 2.3 注意
 
@@ -197,13 +203,7 @@ PPOCRLabel支持三种导出方式：
 
 - 关闭应用程序导出
 
-### 3.4 导出部分识别结果
-
-针对部分难以识别的数据，通过在识别结果的复选框中**取消勾选**相应的标记，其识别结果不会被导出。被取消勾选的识别结果在标记文件 `label.txt` 中的 `difficult` 变量保存为 `True` 。
-
-> *注意：识别结果中的复选框状态仍需用户手动点击确认后才能保留*
-
-### 3.5 数据集划分
+### 3.4 数据集划分
 
 在终端中输入以下命令执行数据集划分脚本：
 
@@ -232,7 +232,7 @@ python gen_ocr_train_val_test.py --trainValTestRatio 6:2:2 --datasetRootPath ../
     | ...
   ```
   
-### 3.6 错误提示
+### 3.5 错误提示
 
 - 如果同时使用whl包安装了paddleocr，其优先级大于通过paddleocr.py调用PaddleOCR类，whl包未更新时会导致程序异常。
 

PPOCRLabel/libs/canvas.py

@@ -627,7 +627,7 @@ class Canvas(QWidget):
         # adaptive BBOX label & index font size
         if self.pixmap:
             h, w = self.pixmap.size().height(), self.pixmap.size().width()
-            fontszie = int(max(h, w) / 48)
+            fontszie = int(max(h, w) / 96)
             for s in self.shapes:
                 s.fontsize = fontszie
         

PPOCRLabel/libs/shape.py

@@ -126,7 +126,7 @@ class Shape(object):
             color = self.select_line_color if self.selected else self.line_color
             pen = QPen(color)
             # Try using integer sizes for smoother drawing(?)
-            pen.setWidth(max(1, int(round(2.0 / self.scale))))
+            # pen.setWidth(max(1, int(round(2.0 / self.scale))))
             painter.setPen(pen)
 
             line_path = QPainterPath()

configs/det/det_mv3_db.yml

@@ -101,7 +101,7 @@ Train:
     drop_last: False
     batch_size_per_card: 16
     num_workers: 8
-    use_shared_memory: False
+    use_shared_memory: True
 
 Eval:
   dataset:
@@ -129,4 +129,4 @@ Eval:
     drop_last: False
     batch_size_per_card: 1 # must be 1
     num_workers: 8
-    use_shared_memory: False
+    use_shared_memory: True

configs/table/table_master.yml

@@ -104,7 +104,7 @@ Train:
 Eval:
   dataset:
     name: PubTabDataSet
-    data_dir: train_data/table/pubtabnet/train/
+    data_dir: train_data/table/pubtabnet/val/
     label_file_list: [train_data/table/pubtabnet/PubTabNet_2.0.0_val.jsonl]
     transforms:
       - DecodeImage:

doc/doc_ch/algorithm_overview.md

@@ -91,12 +91,9 @@
 |SVTR|SVTR-Tiny| 89.25% | rec_svtr_tiny_none_ctc_en | [训练模型](https://paddleocr.bj.bcebos.com/PP-OCRv3/chinese/rec_svtr_tiny_none_ctc_en_train.tar) |
 |ViTSTR|ViTSTR| 79.82% | rec_vitstr_none_ce | [训练模型](https://paddleocr.bj.bcebos.com/rec_vitstr_none_ce_train.tar) |
 |ABINet|Resnet45| 90.75% | rec_r45_abinet | [训练模型](https://paddleocr.bj.bcebos.com/rec_r45_abinet_train.tar) |
-<<<<<<< HEAD
-|RobustScanner|ResNet31V2| 87.77% | rec_r31_robustscanner | coming soon |
-=======
 |SPIN|ResNet32| 90.00% | rec_r32_gaspin_bilstm_att | coming soon |
+|RobustScanner|ResNet31V2| 87.77% | rec_r31_robustscanner | coming soon |
 
->>>>>>> 1696b36bdb4152138ed5cb08a357df8fe03dc067
 
 <a name="2"></a>
 

doc/doc_en/algorithm_overview_en.md

@@ -68,11 +68,8 @@ Supported text recognition algorithms (Click the link to get the tutorial):
 - [x]  [SVTR](./algorithm_rec_svtr_en.md)
 - [x]  [ViTSTR](./algorithm_rec_vitstr_en.md)
 - [x]  [ABINet](./algorithm_rec_abinet_en.md)
-<<<<<<< HEAD
-- [x]  [RobustScanner](./algorithm_rec_robustscanner_en.md)
-=======
 - [x]  [SPIN](./algorithm_rec_spin_en.md)
->>>>>>> 1696b36bdb4152138ed5cb08a357df8fe03dc067
+- [x]  [RobustScanner](./algorithm_rec_robustscanner_en.md)
 
 Refer to [DTRB](https://arxiv.org/abs/1904.01906), the training and evaluation result of these above text recognition (using MJSynth and SynthText for training, evaluate on IIIT, SVT, IC03, IC13, IC15, SVTP, CUTE) is as follow:
 

test_tipc/configs/ch_PP-OCRv2_det/train_infer_python.txt

@@ -13,7 +13,7 @@ train_infer_img_dir:./train_data/icdar2015/text_localization/ch4_test_images/
 null:null
 ##
 trainer:norm_train
-norm_train:tools/train.py -c configs/det/ch_PP-OCRv2/ch_PP-OCRv2_det_cml.yml -o 
+norm_train:tools/train.py -c configs/det/ch_PP-OCRv2/ch_PP-OCRv2_det_cml.yml -o Global.print_batch_step=1 Train.loader.shuffle=false
 pact_train:null
 fpgm_train:null
 distill_train:null
@@ -51,3 +51,9 @@ null:null
 null:null
 ===========================infer_benchmark_params==========================
 random_infer_input:[{float32,[3,640,640]}];[{float32,[3,960,960]}]
+===========================train_benchmark_params==========================
+batch_size:8
+fp_items:fp32|fp16
+epoch:2
+--profiler_options:batch_range=[10,20];state=GPU;tracer_option=Default;profile_path=model.profile
+flags:FLAGS_eager_delete_tensor_gb=0.0;FLAGS_fraction_of_gpu_memory_to_use=0.98;FLAGS_conv_workspace_size_limit=4096

test_tipc/configs/ch_PP-OCRv2_rec/ch_PP-OCRv2_rec_distillation.yml

@@ -6,7 +6,7 @@ Global:
   print_batch_step: 10
   save_model_dir: ./output/rec_pp-OCRv2_distillation
   save_epoch_step: 3
-  eval_batch_step: [0, 2000]
+  eval_batch_step: [0, 200000]
   cal_metric_during_train: true
   pretrained_model:
   checkpoints:
@@ -114,7 +114,7 @@ Train:
     name: SimpleDataSet
     data_dir: ./train_data/ic15_data/
     label_file_list:
-    - ./train_data/ic15_data/rec_gt_train.txt
+    - ./train_data/ic15_data/rec_gt_train4w.txt
     transforms:
     - DecodeImage:
         img_mode: BGR

test_tipc/configs/ch_PP-OCRv2_rec/train_infer_python.txt

@@ -13,7 +13,7 @@ train_infer_img_dir:./inference/rec_inference
 null:null
 ##
 trainer:norm_train
-norm_train:tools/train.py -c test_tipc/configs/ch_PP-OCRv2_rec/ch_PP-OCRv2_rec_distillation.yml -o 
+norm_train:tools/train.py -c test_tipc/configs/ch_PP-OCRv2_rec/ch_PP-OCRv2_rec_distillation.yml -o Global.print_batch_step=4 Train.loader.shuffle=false
 pact_train:null
 fpgm_train:null
 distill_train:null
@@ -51,3 +51,9 @@ null:null
 null:null
 ===========================infer_benchmark_params==========================
 random_infer_input:[{float32,[3,32,320]}]
+===========================train_benchmark_params==========================
+batch_size:64
+fp_items:fp32|fp16
+epoch:1
+--profiler_options:batch_range=[10,20];state=GPU;tracer_option=Default;profile_path=model.profile
+flags:FLAGS_eager_delete_tensor_gb=0.0;FLAGS_fraction_of_gpu_memory_to_use=0.98;FLAGS_conv_workspace_size_limit=4096

test_tipc/configs/ch_PP-OCRv3_det/train_infer_python.txt

@@ -13,7 +13,7 @@ train_infer_img_dir:./train_data/icdar2015/text_localization/ch4_test_images/
 null:null
 ##
 trainer:norm_train
-norm_train:tools/train.py -c configs/det/ch_PP-OCRv3/ch_PP-OCRv3_det_cml.yml -o 
+norm_train:tools/train.py -c configs/det/ch_PP-OCRv3/ch_PP-OCRv3_det_cml.yml -o Global.print_batch_step=1 Train.loader.shuffle=false Global.eval_batch_step=[4000,400]
 pact_train:null
 fpgm_train:null
 distill_train:null
@@ -51,3 +51,9 @@ null:null
 null:null
 ===========================infer_benchmark_params==========================
 random_infer_input:[{float32,[3,640,640]}];[{float32,[3,960,960]}]
+===========================train_benchmark_params==========================
+batch_size:8
+fp_items:fp32|fp16
+epoch:2
+--profiler_options:batch_range=[10,20];state=GPU;tracer_option=Default;profile_path=model.profile
+flags:FLAGS_eager_delete_tensor_gb=0.0;FLAGS_fraction_of_gpu_memory_to_use=0.98;FLAGS_conv_workspace_size_limit=4096

test_tipc/configs/ch_PP-OCRv3_rec/ch_PP-OCRv3_rec_distillation.yml

@@ -153,7 +153,7 @@ Train:
     data_dir: ./train_data/ic15_data/
     ext_op_transform_idx: 1
     label_file_list:
-    - ./train_data/ic15_data/rec_gt_train_lite.txt
+    - ./train_data/ic15_data/rec_gt_train4w.txt
     transforms:
     - DecodeImage:
         img_mode: BGR
@@ -183,7 +183,7 @@ Eval:
     name: SimpleDataSet
     data_dir: ./train_data/ic15_data
     label_file_list:
-    - ./train_data/ic15_data/rec_gt_test_lite.txt
+    - ./train_data/ic15_data/rec_gt_test.txt
     transforms:
     - DecodeImage:
         img_mode: BGR

test_tipc/configs/ch_PP-OCRv3_rec/train_infer_python.txt

@@ -13,7 +13,7 @@ train_infer_img_dir:./inference/rec_inference
 null:null
 ##
 trainer:norm_train
-norm_train:tools/train.py -c test_tipc/configs/ch_PP-OCRv3_rec/ch_PP-OCRv3_rec_distillation.yml -o 
+norm_train:tools/train.py -c test_tipc/configs/ch_PP-OCRv3_rec/ch_PP-OCRv3_rec_distillation.yml -o Global.print_batch_step=1 Train.loader.shuffle=false
 pact_train:null
 fpgm_train:null
 distill_train:null
@@ -51,3 +51,9 @@ null:null
 null:null
 ===========================infer_benchmark_params==========================
 random_infer_input:[{float32,[3,48,320]}]
+===========================train_benchmark_params==========================
+batch_size:128
+fp_items:fp32|fp16
+epoch:1
+--profiler_options:batch_range=[10,20];state=GPU;tracer_option=Default;profile_path=model.profile
+flags:FLAGS_eager_delete_tensor_gb=0.0;FLAGS_fraction_of_gpu_memory_to_use=0.98;FLAGS_conv_workspace_size_limit=4096

test_tipc/configs/ch_ppocr_mobile_v2.0_det/train_infer_python.txt

@@ -13,7 +13,7 @@ train_infer_img_dir:./train_data/icdar2015/text_localization/ch4_test_images/
 null:null
 ##
 trainer:norm_train
-norm_train:tools/train.py -c configs/det/ch_ppocr_v2.0/ch_det_mv3_db_v2.0.yml -o Global.pretrained_model=./pretrain_models/MobileNetV3_large_x0_5_pretrained
+norm_train:tools/train.py -c configs/det/ch_ppocr_v2.0/ch_det_mv3_db_v2.0.yml -o Global.pretrained_model=./pretrain_models/MobileNetV3_large_x0_5_pretrained Global.print_batch_step=1 Train.loader.shuffle=false
 pact_train:null
 fpgm_train:null
 distill_train:null
@@ -50,4 +50,10 @@ null:null
 --benchmark:True
 null:null
 ===========================infer_benchmark_params==========================
-random_infer_input:[{float32,[3,640,640]}];[{float32,[3,960,960]}]
\ No newline at end of file
+random_infer_input:[{float32,[3,640,640]}];[{float32,[3,960,960]}]
+===========================train_benchmark_params==========================
+batch_size:8
+fp_items:fp32|fp16
+epoch:2
+--profiler_options:batch_range=[10,20];state=GPU;tracer_option=Default;profile_path=model.profile
+flags:FLAGS_eager_delete_tensor_gb=0.0;FLAGS_fraction_of_gpu_memory_to_use=0.98;FLAGS_conv_workspace_size_limit=4096

test_tipc/configs/ch_ppocr_mobile_v2.0_rec/train_infer_python.txt

@@ -13,7 +13,7 @@ train_infer_img_dir:./inference/rec_inference
 null:null
 ##
 trainer:norm_train
-norm_train:tools/train.py -c configs/rec/rec_icdar15_train.yml -o
+norm_train:tools/train.py -c configs/rec/rec_icdar15_train.yml -o Global.print_batch_step=4 Train.loader.shuffle=false
 pact_train:null
 fpgm_train:null
 distill_train:null
@@ -51,3 +51,9 @@ inference:tools/infer/predict_rec.py
 null:null
 ===========================infer_benchmark_params==========================
 random_infer_input:[{float32,[3,32,100]}]
+===========================train_benchmark_params==========================
+batch_size:256
+fp_items:fp32|fp16
+epoch:3
+--profiler_options:batch_range=[10,20];state=GPU;tracer_option=Default;profile_path=model.profile
+flags:FLAGS_eager_delete_tensor_gb=0.0;FLAGS_fraction_of_gpu_memory_to_use=0.98;FLAGS_conv_workspace_size_limit=4096

test_tipc/configs/ch_ppocr_server_v2.0_det/det_r50_vd_db.yml

@@ -2,13 +2,13 @@ Global:
   use_gpu: false
   epoch_num: 5
   log_smooth_window: 20
-  print_batch_step: 1
+  print_batch_step: 2
   save_model_dir: ./output/db_mv3/
   save_epoch_step: 1200
   # evaluation is run every 2000 iterations
-  eval_batch_step: [0, 400]
+  eval_batch_step: [0, 30000]
   cal_metric_during_train: False
-  pretrained_model: ./pretrain_models/MobileNetV3_large_x0_5_pretrained
+  pretrained_model:
   checkpoints:
   save_inference_dir:
   use_visualdl: False

test_tipc/configs/ch_ppocr_server_v2.0_det/train_infer_python.txt

@@ -13,7 +13,7 @@ train_infer_img_dir:./train_data/icdar2015/text_localization/ch4_test_images/
 null:null
 ##
 trainer:norm_train
-norm_train:tools/train.py -c test_tipc/configs/ch_ppocr_server_v2.0_det/det_r50_vd_db.yml -o 
+norm_train:tools/train.py -c test_tipc/configs/ch_ppocr_server_v2.0_det/det_r50_vd_db.yml -o
 quant_train:null
 fpgm_train:null
 distill_train:null
@@ -50,4 +50,10 @@ inference:tools/infer/predict_det.py
 --benchmark:True
 null:null
 ===========================infer_benchmark_params==========================
-random_infer_input:[{float32,[3,640,640]}];[{float32,[3,960,960]}]
\ No newline at end of file
+random_infer_input:[{float32,[3,640,640]}];[{float32,[3,960,960]}]
+===========================train_benchmark_params==========================
+batch_size:8
+fp_items:fp32|fp16
+epoch:2
+--profiler_options:batch_range=[10,20];state=GPU;tracer_option=Default;profile_path=model.profile
+flags:FLAGS_eager_delete_tensor_gb=0.0;FLAGS_fraction_of_gpu_memory_to_use=0.98;FLAGS_conv_workspace_size_limit=4096

test_tipc/configs/ch_ppocr_server_v2.0_rec/train_infer_python.txt

@@ -13,7 +13,7 @@ train_infer_img_dir:./inference/rec_inference
 null:null
 ##
 trainer:norm_train
-norm_train:tools/train.py -c test_tipc/configs/ch_ppocr_server_v2.0_rec/rec_icdar15_train.yml -o
+norm_train:tools/train.py -c test_tipc/configs/ch_ppocr_server_v2.0_rec/rec_icdar15_train.yml -o Global.print_batch_step=4 Train.loader.shuffle=false
 pact_train:null
 fpgm_train:null
 distill_train:null
@@ -51,3 +51,9 @@ inference:tools/infer/predict_rec.py
 null:null
 ===========================infer_benchmark_params==========================
 random_infer_input:[{float32,[3,32,100]}]
+===========================train_benchmark_params==========================
+batch_size:256
+fp_items:fp32|fp16
+epoch:2
+--profiler_options:batch_range=[10,20];state=GPU;tracer_option=Default;profile_path=model.profile
+flags:FLAGS_eager_delete_tensor_gb=0.0;FLAGS_fraction_of_gpu_memory_to_use=0.98;FLAGS_conv_workspace_size_limit=4096

test_tipc/configs/det_mv3_db_v2_0/train_infer_python.txt

@@ -13,7 +13,7 @@ train_infer_img_dir:./train_data/icdar2015/text_localization/ch4_test_images/
 null:null
 ##
 trainer:norm_train
-norm_train:tools/train.py -c configs/det/det_mv3_db.yml -o Global.pretrained_model=./pretrain_models/MobileNetV3_large_x0_5_pretrained
+norm_train:tools/train.py -c configs/det/det_mv3_db.yml -o Global.pretrained_model=./pretrain_models/MobileNetV3_large_x0_5_pretrained Global.print_batch_step=1 Train.loader.shuffle=false
 pact_train:null
 fpgm_train:null
 distill_train:null
@@ -52,8 +52,8 @@ null:null
 ===========================infer_benchmark_params==========================
 random_infer_input:[{float32,[3,640,640]}];[{float32,[3,960,960]}]
 ===========================train_benchmark_params==========================
-batch_size:8|16
+batch_size:16
 fp_items:fp32|fp16
-epoch:15
+epoch:4
 --profiler_options:batch_range=[10,20];state=GPU;tracer_option=Default;profile_path=model.profile
-flags:FLAGS_eager_delete_tensor_gb=0.0;FLAGS_fraction_of_gpu_memory_to_use=0.98;FLAGS_conv_workspace_size_limit=4096
+flags:FLAGS_eager_delete_tensor_gb=0.0;FLAGS_fraction_of_gpu_memory_to_use=0.98;FLAGS_conv_workspace_size_limit=4096
\ No newline at end of file

test_tipc/configs/det_r50_db_v2.0/train_infer_python.txt

@@ -13,7 +13,7 @@ train_infer_img_dir:./train_data/icdar2015/text_localization/ch4_test_images/
 null:null
 ##
 trainer:norm_train
-norm_train:tools/train.py -c configs/det/det_r50_vd_db.yml -o 
+norm_train:tools/train.py -c configs/det/det_r50_vd_db.yml -o Global.print_batch_step=1 Train.loader.shuffle=false
 quant_export:null
 fpgm_export:null
 distill_train:null
@@ -50,4 +50,10 @@ inference:tools/infer/predict_det.py
 --benchmark:True
 null:null
 ===========================infer_benchmark_params==========================
-random_infer_input:[{float32,[3,640,640]}];[{float32,[3,960,960]}]
\ No newline at end of file
+random_infer_input:[{float32,[3,640,640]}];[{float32,[3,960,960]}]
+===========================train_benchmark_params==========================
+batch_size:8
+fp_items:fp32|fp16
+epoch:2
+--profiler_options:batch_range=[10,20];state=GPU;tracer_option=Default;profile_path=model.profile
+flags:FLAGS_eager_delete_tensor_gb=0.0;FLAGS_fraction_of_gpu_memory_to_use=0.98;FLAGS_conv_workspace_size_limit=4096
\ No newline at end of file

test_tipc/configs/det_r50_dcn_fce_ctw_v2.0/det_r50_vd_dcn_fce_ctw.yml

+Global:
+  use_gpu: true
+  epoch_num: 1500
+  log_smooth_window: 20
+  print_batch_step: 20
+  save_model_dir: ./output/det_r50_dcn_fce_ctw/
+  save_epoch_step: 100
+  # evaluation is run every 835 iterations
+  eval_batch_step: [0, 4000]
+  cal_metric_during_train: False
+  pretrained_model: ./pretrain_models/ResNet50_vd_ssld_pretrained 
+  checkpoints: 
+  save_inference_dir: 
+  use_visualdl: False
+  infer_img: doc/imgs_en/img_10.jpg
+  save_res_path: ./output/det_fce/predicts_fce.txt
+
+
+Architecture:
+  model_type: det
+  algorithm: FCE
+  Transform:
+  Backbone:
+    name: ResNet_vd
+    layers: 50
+    dcn_stage: [False, True, True, True]
+    out_indices: [1,2,3]
+  Neck:
+    name: FCEFPN
+    out_channels: 256
+    has_extra_convs: False
+    extra_stage: 0
+  Head:
+    name: FCEHead
+    fourier_degree: 5
+Loss:
+  name: FCELoss
+  fourier_degree: 5
+  num_sample: 50
+  
+Optimizer:
+  name: Adam
+  beta1: 0.9
+  beta2: 0.999
+  lr:
+    learning_rate: 0.0001
+  regularizer:
+    name: 'L2'
+    factor: 0
+
+PostProcess:
+  name: FCEPostProcess
+  scales: [8, 16, 32]
+  alpha: 1.0
+  beta: 1.0
+  fourier_degree: 5
+  box_type: 'poly'
+
+Metric:
+  name: DetFCEMetric
+  main_indicator: hmean
+
+Train:
+  dataset:
+    name: SimpleDataSet
+    data_dir: ./train_data/icdar2015/text_localization/
+    label_file_list:
+      - ./train_data/icdar2015/text_localization/train_icdar2015_label.txt
+    transforms:
+      - DecodeImage: # load image
+          img_mode: BGR
+          channel_first: False
+          ignore_orientation: True
+      - DetLabelEncode: # Class handling label
+      - ColorJitter: 
+          brightness: 0.142
+          saturation: 0.5
+          contrast: 0.5
+      - RandomScaling: 
+      - RandomCropFlip:
+          crop_ratio: 0.5
+      - RandomCropPolyInstances:
+          crop_ratio: 0.8
+          min_side_ratio: 0.3
+      - RandomRotatePolyInstances:
+          rotate_ratio: 0.5
+          max_angle: 30
+          pad_with_fixed_color: False
+      - SquareResizePad:
+          target_size: 800
+          pad_ratio: 0.6
+      - IaaAugment:
+          augmenter_args:
+            - { 'type': Fliplr, 'args': { 'p': 0.5 } }
+      - FCENetTargets:
+          fourier_degree: 5
+      - NormalizeImage:
+          scale: 1./255.
+          mean: [0.485, 0.456, 0.406]
+          std: [0.229, 0.224, 0.225]
+          order: 'hwc'
+      - ToCHWImage:
+      - KeepKeys:
+          keep_keys: ['image', 'p3_maps', 'p4_maps', 'p5_maps'] # dataloader will return list in this order
+  loader:
+    shuffle: True
+    drop_last: False
+    batch_size_per_card: 6
+    num_workers: 8
+
+Eval:
+  dataset:
+    name: SimpleDataSet
+    data_dir: ./train_data/icdar2015/text_localization/
+    label_file_list:
+      - ./train_data/icdar2015/text_localization/test_icdar2015_label.txt
+    transforms:
+      - DecodeImage: # load image
+          img_mode: BGR
+          channel_first: False
+          ignore_orientation: True
+      - DetLabelEncode: # Class handling label
+      - DetResizeForTest:
+          limit_type: 'min'
+          limit_side_len: 736
+      - NormalizeImage:
+          scale: 1./255.
+          mean: [0.485, 0.456, 0.406]
+          std: [0.229, 0.224, 0.225]
+          order: 'hwc'
+      - Pad: 
+      - ToCHWImage:
+      - KeepKeys:
+          keep_keys: ['image', 'shape', 'polys', 'ignore_tags']
+  loader:
+    shuffle: False
+    drop_last: False
+    batch_size_per_card: 1 # must be 1
+    num_workers: 2
\ No newline at end of file

test_tipc/configs/det_r50_dcn_fce_ctw_v2.0/train_infer_python.txt

+===========================train_params===========================
+model_name:det_r50_dcn_fce_ctw_v2.0
+python:python3.7
+gpu_list:0
+Global.use_gpu:True|True
+Global.auto_cast:fp32
+Global.epoch_num:lite_train_lite_infer=1|whole_train_whole_infer=500
+Global.save_model_dir:./output/
+Train.loader.batch_size_per_card:lite_train_lite_infer=2|whole_train_whole_infer=4
+Global.pretrained_model:null
+train_model_name:latest
+train_infer_img_dir:./train_data/icdar2015/text_localization/ch4_test_images/
+null:null
+##
+trainer:norm_train
+norm_train:tools/train.py -c test_tipc/configs/det_r50_dcn_fce_ctw_v2.0/det_r50_vd_dcn_fce_ctw.yml -o Global.print_batch_step=1 Train.loader.shuffle=false
+pact_train:null
+fpgm_train:null
+distill_train:null
+null:null
+null:null
+##
+===========================eval_params=========================== 
+eval:null
+null:null
+##
+===========================infer_params===========================
+Global.save_inference_dir:./output/
+Global.checkpoints:
+norm_export:tools/export_model.py -c test_tipc/configs/det_r50_dcn_fce_ctw_v2.0/det_r50_vd_dcn_fce_ctw.yml -o 
+quant_export:null 
+fpgm_export:null
+distill_export:null
+export1:null
+export2:null
+##
+train_model:./inference/det_r50_dcn_fce_ctw_v2.0_train/best_accuracy
+infer_export:tools/export_model.py -c test_tipc/configs/det_r50_dcn_fce_ctw_v2.0/det_r50_vd_dcn_fce_ctw.yml  -o
+infer_quant:False
+inference:tools/infer/predict_det.py
+--use_gpu:True|False
+--enable_mkldnn:False
+--cpu_threads:6
+--rec_batch_num:1
+--use_tensorrt:False
+--precision:fp32
+--det_model_dir:
+--image_dir:./inference/ch_det_data_50/all-sum-510/
+--save_log_path:null
+--benchmark:True
+--det_algorithm:FCE
+===========================infer_benchmark_params==========================
+random_infer_input:[{float32,[3,640,640]}];[{float32,[3,960,960]}]
+===========================train_benchmark_params==========================
+batch_size:6
+fp_items:fp32|fp16
+epoch:1
+--profiler_options:batch_range=[10,20];state=GPU;tracer_option=Default;profile_path=model.profile
+flags:FLAGS_eager_delete_tensor_gb=0.0;FLAGS_fraction_of_gpu_memory_to_use=0.98;FLAGS_conv_workspace_size_limit=4096
\ No newline at end of file

test_tipc/configs/det_r50_vd_east_v2_0/det_r50_vd_east.yml

@@ -20,7 +20,7 @@ Architecture:
   algorithm: EAST
   Transform:
   Backbone:
-    name: ResNet
+    name: ResNet_vd
     layers: 50
   Neck:
     name: EASTFPN

test_tipc/configs/det_r50_vd_east_v2_0/train_infer_python.txt

@@ -13,7 +13,7 @@ train_infer_img_dir:./train_data/icdar2015/text_localization/ch4_test_images/
 null:null
 ##
 trainer:norm_train
-norm_train:tools/train.py -c test_tipc/configs/det_r50_vd_east_v2_0/det_r50_vd_east.yml -o 
+norm_train:tools/train.py -c test_tipc/configs/det_r50_vd_east_v2_0/det_r50_vd_east.yml -o Global.pretrained_model=pretrain_models/det_r50_vd_east_v2.0_train/best_accuracy.pdparams Global.print_batch_step=1 Train.loader.shuffle=false
 pact_train:null
 fpgm_train:null
 distill_train:null
@@ -55,4 +55,5 @@ random_infer_input:[{float32,[3,640,640]}];[{float32,[3,960,960]}]
 batch_size:8
 fp_items:fp32|fp16
 epoch:2
---profiler_options:batch_range=[10,20];state=GPU;tracer_option=Default;profile_path=model.profile
\ No newline at end of file
+--profiler_options:batch_range=[10,20];state=GPU;tracer_option=Default;profile_path=model.profile
+flags:FLAGS_eager_delete_tensor_gb=0.0;FLAGS_fraction_of_gpu_memory_to_use=0.98;FLAGS_conv_workspace_size_limit=4096
\ No newline at end of file

test_tipc/configs/det_r50_vd_pse_v2_0/det_r50_vd_pse.yml

@@ -8,7 +8,7 @@ Global:
   # evaluation is run every 125 iterations
   eval_batch_step: [ 0,1000 ]
   cal_metric_during_train: False
-  pretrained_model: 
+  pretrained_model: ./pretrain_models/ResNet50_vd_ssld_pretrained
   checkpoints: #./output/det_r50_vd_pse_batch8_ColorJitter/best_accuracy
   save_inference_dir:
   use_visualdl: False
@@ -20,7 +20,7 @@ Architecture:
   algorithm: PSE
   Transform:
   Backbone:
-    name: ResNet
+    name: ResNet_vd
     layers: 50
   Neck:
     name: FPN

test_tipc/configs/det_r50_vd_pse_v2_0/train_infer_python.txt

@@ -13,7 +13,7 @@ train_infer_img_dir:./train_data/icdar2015/text_localization/ch4_test_images/
 null:null
 ##
 trainer:norm_train
-norm_train:tools/train.py -c test_tipc/configs/det_r50_vd_pse_v2_0/det_r50_vd_pse.yml -o 
+norm_train:tools/train.py -c test_tipc/configs/det_r50_vd_pse_v2_0/det_r50_vd_pse.yml -o Global.print_batch_step=1 Train.loader.shuffle=false
 pact_train:null
 fpgm_train:null
 distill_train:null
@@ -54,5 +54,6 @@ random_infer_input:[{float32,[3,640,640]}];[{float32,[3,960,960]}]
 ===========================train_benchmark_params==========================
 batch_size:8
 fp_items:fp32|fp16
-epoch:10
+epoch:2
 --profiler_options:batch_range=[10,20];state=GPU;tracer_option=Default;profile_path=model.profile
+flags:FLAGS_eager_delete_tensor_gb=0.0;FLAGS_fraction_of_gpu_memory_to_use=0.98;FLAGS_conv_workspace_size_limit=4096
\ No newline at end of file

test_tipc/configs/en_table_structure/table_mv3.yml

@@ -6,7 +6,7 @@ Global:
   save_model_dir: ./output/table_mv3/
   save_epoch_step: 3
   # evaluation is run every 400 iterations after the 0th iteration
-  eval_batch_step: [0, 400]
+  eval_batch_step: [0, 40000]
   cal_metric_during_train: True
   pretrained_model:
   checkpoints:

test_tipc/configs/en_table_structure/train_infer_python.txt

@@ -13,7 +13,7 @@ train_infer_img_dir:./ppstructure/docs/table/table.jpg
 null:null
 ##
 trainer:norm_train
-norm_train:tools/train.py -c test_tipc/configs/en_table_structure/table_mv3.yml -o 
+norm_train:tools/train.py -c test_tipc/configs/en_table_structure/table_mv3.yml -o Global.print_batch_step=1 Train.loader.shuffle=false
 pact_train:null
 fpgm_train:null
 distill_train:null
@@ -27,7 +27,7 @@ null:null
 ===========================infer_params===========================
 Global.save_inference_dir:./output/
 Global.checkpoints:
-norm_export:tools/export_model.py -c test_tipc/configs/en_table_structure/table_mv3.yml -o 
+norm_export:tools/export_model.py -c test_tipc/configs/en_table_structure/table_mv3.yml -o
 quant_export:
 fpgm_export: 
 distill_export:null
@@ -51,3 +51,9 @@ null:null
 null:null
 ===========================infer_benchmark_params==========================
 random_infer_input:[{float32,[3,488,488]}]
+===========================train_benchmark_params==========================
+batch_size:32
+fp_items:fp32|fp16
+epoch:1
+--profiler_options:batch_range=[10,20];state=GPU;tracer_option=Default;profile_path=model.profile
+flags:FLAGS_eager_delete_tensor_gb=0.0;FLAGS_fraction_of_gpu_memory_to_use=0.98;FLAGS_conv_workspace_size_limit=4096

test_tipc/configs/rec_mv3_none_bilstm_ctc_v2.0/rec_icdar15_train.yml

@@ -6,7 +6,7 @@ Global:
   save_model_dir: ./output/rec/mv3_none_bilstm_ctc/
   save_epoch_step: 3
   # evaluation is run every 2000 iterations
-  eval_batch_step: [0, 2000]
+  eval_batch_step: [0, 20000]
   cal_metric_during_train: True
   pretrained_model:
   checkpoints:

test_tipc/configs/rec_mv3_none_bilstm_ctc_v2.0/train_infer_python.txt

@@ -13,7 +13,7 @@ train_infer_img_dir:./inference/rec_inference
 null:null
 ##
 trainer:norm_train
-norm_train:tools/train.py -c test_tipc/configs/rec_mv3_none_bilstm_ctc_v2.0/rec_icdar15_train.yml -o
+norm_train:tools/train.py -c test_tipc/configs/rec_mv3_none_bilstm_ctc_v2.0/rec_icdar15_train.yml -o Global.print_batch_step=4 Train.loader.shuffle=false
 pact_train:null
 fpgm_train:null
 distill_train:null
@@ -50,4 +50,10 @@ inference:tools/infer/predict_rec.py --rec_char_dict_path=./ppocr/utils/ic15_dic
 --benchmark:True
 null:null
 ===========================infer_benchmark_params==========================
-random_infer_input:[{float32,[3,32,100]}]
\ No newline at end of file
+random_infer_input:[{float32,[3,32,100]}]
+===========================train_benchmark_params==========================
+batch_size:256
+fp_items:fp32|fp16
+epoch:4
+--profiler_options:batch_range=[10,20];state=GPU;tracer_option=Default;profile_path=model.profile
+flags:FLAGS_eager_delete_tensor_gb=0.0;FLAGS_fraction_of_gpu_memory_to_use=0.98;FLAGS_conv_workspace_size_limit=4096

test_tipc/docs/benchmark_train.md

@@ -51,3 +51,25 @@ train_log/
     ├── PaddleOCR_det_mv3_db_v2_0_bs8_fp32_SingleP_DP_N1C1_log
     └── PaddleOCR_det_mv3_db_v2_0_bs8_fp32_SingleP_DP_N1C4_log
 ```
+## 3. 各模型单卡性能数据一览
+
+*注：本节中的速度指标均使用单卡（1块Nvidia V100 16G GPU）测得。通常情况下。
+
+
+|模型名称|配置文件|大数据集 float32 fps |小数据集 float32 fps |diff |大数据集 float16 fps|小数据集 float16 fps| diff | 大数据集大小 | 小数据集大小 |
+|:-:|:-:|:-:|:-:|:-:|:-:|:-:|:-:|:-:|:-:|
+| ch_ppocr_mobile_v2.0_det |[config](../configs/ch_ppocr_mobile_v2.0_det/train_infer_python.txt) | 53.836 | 53.343 / 53.914 / 52.785 |0.020940758 | 45.574 | 45.57 / 46.292 / 46.213 | 0.015596647 | 10,000| 2,000|
+| ch_ppocr_mobile_v2.0_rec |[config](../configs/ch_ppocr_mobile_v2.0_rec/train_infer_python.txt) | 2083.311 | 2043.194	/ 2066.372 / 2093.317 |0.023944295 | 2153.261 | 2167.561 /	2165.726 /	2155.614| 0.005511725 | 600,000| 160,000|
+| ch_ppocr_server_v2.0_det |[config](../configs/ch_ppocr_server_v2.0_det/train_infer_python.txt) | 20.716 | 20.739 /	20.807 /	20.755 |0.003268131 | 20.592 | 20.498 /	20.993 /	20.75| 0.023579288 | 10,000| 2,000|
+| ch_ppocr_server_v2.0_rec |[config](../configs/ch_ppocr_server_v2.0_rec/train_infer_python.txt) | 528.56 | 528.386 /	528.991 /	528.391 |0.001143687 | 1189.788 | 1190.007 /	1176.332 /	1192.084| 0.013213834 |  600,000| 160,000|
+| ch_PP-OCRv2_det	 |[config](../configs/ch_PP-OCRv2_det/train_infer_python.txt) | 13.87 | 13.386 /	13.529 /	13.428 |0.010569887 | 17.847 | 17.746 /	17.908 /	17.96| 0.011915367 | 10,000| 2,000|
+| ch_PP-OCRv2_rec	 |[config](../configs/ch_PP-OCRv2_rec/train_infer_python.txt) | 109.248 | 106.32 /	106.318 /	108.587 |0.020895687 | 117.491 | 117.62 /	117.757 /	117.726| 0.001163413 | 140,000| 40,000|
+| det_mv3_db_v2.0	 |[config](../configs/det_mv3_db_v2_0/train_infer_python.txt) | 61.802 | 62.078 /	61.802 /	62.008 |0.00444602 | 82.947 | 84.294 /	84.457 /	84.005| 0.005351836 | 10,000| 2,000|
+| det_r50_vd_db_v2.0	 |[config](../configs/det_r50_vd_db_v2.0/train_infer_python.txt) | 29.955 | 29.092 /	29.31 /	28.844 |0.015899011 | 51.097 |50.367 /	50.879 /	50.227| 0.012814717 | 10,000| 2,000|
+| det_r50_vd_east_v2.0	 |[config](../configs/det_r50_vd_east_v2.0/train_infer_python.txt) | 42.485 | 42.624 /	42.663 /	42.561 |0.00239083 | 67.61 |67.825/ 	68.299/ 	68.51| 0.00999854 | 10,000| 2,000|
+| det_r50_vd_pse_v2.0	 |[config](../configs/det_r50_vd_pse_v2.0/train_infer_python.txt) | 16.455 | 16.517 / 16.555 /	16.353 |0.012201752 | 27.02 |27.288 /	27.152 /	27.408| 0.009340339 | 10,000| 2,000|
+| rec_mv3_none_bilstm_ctc_v2.0	 |[config](../configs/rec_mv3_none_bilstm_ctc_v2.0/train_infer_python.txt) | 2288.358 | 2291.906 /	2293.725 /	2290.05 |0.001602197 | 2336.17 |2327.042 /	2328.093 /	2344.915| 0.007622025 | 600,000| 160,000|
+| PP-Structure-table	 |[config](../configs/en_table_structure/train_infer_python.txt) | 14.151 | 14.077 /	14.23 /	14.25 |0.012140351 | 16.285 | 16.595 /	16.878 /	16.531 | 0.020559308 | 20,000| 5,000|
+| det_r50_dcn_fce_ctw_v2.0	 |[config](../configs/det_r50_dcn_fce_ctw_v2.0/train_infer_python.txt) | 14.057 | 14.029 /	14.02 /	14.014 |0.001069214 | 18.298 |18.411 /	18.376 /	18.331| 0.004345228 | 10,000| 2,000|
+| ch_PP-OCRv3_det	 |[config](../configs/ch_PP-OCRv3_det/train_infer_python.txt) | 8.622 | 8.431 /	8.423 /	8.479|0.006604552 | 14.203 |14.346	14.468	14.23| 0.016450097 | 10,000| 2,000|
+| ch_PP-OCRv3_rec	 |[config](../configs/ch_PP-OCRv3_rec/train_infer_python.txt) | 73.627 | 72.46 /	73.575 /	73.704|0.016878324 | | |  | 160,000| 40,000|
\ No newline at end of file

test_tipc/prepare.sh

@@ -22,27 +22,79 @@ trainer_list=$(func_parser_value "${lines[14]}")
 
 if [ ${MODE} = "benchmark_train" ];then
     pip install -r requirements.txt
-    if [[ ${model_name} =~ "det_mv3_db_v2_0" || ${model_name} =~ "det_r50_vd_pse_v2_0" || ${model_name} =~ "det_r18_db_v2_0" ]];then
-        rm -rf ./train_data/icdar2015
+    if [[ ${model_name} =~ "ch_ppocr_mobile_v2.0_det" || ${model_name} =~ "det_mv3_db_v2_0" ]];then
         wget -nc -P  ./pretrain_models/ https://paddleocr.bj.bcebos.com/pretrained/MobileNetV3_large_x0_5_pretrained.pdparams  --no-check-certificate
-        wget -nc -P ./train_data/ https://paddleocr.bj.bcebos.com/dygraph_v2.0/test/icdar2015.tar --no-check-certificate
-        cd ./train_data/ && tar xf icdar2015.tar && cd ../
+        rm -rf ./train_data/icdar2015
+        wget -nc -P ./train_data/ https://paddleocr.bj.bcebos.com/dataset/icdar2015_benckmark.tar --no-check-certificate
+        cd ./train_data/ && tar xf icdar2015_benckmark.tar
+        ln -s ./icdar2015_benckmark ./icdar2015
+        cd ../
+    fi
+    if [[ ${model_name} =~ "ch_ppocr_server_v2.0_det" || ${model_name} =~ "ch_PP-OCRv3_det" ]];then
+        rm -rf ./train_data/icdar2015
+        wget -nc -P ./train_data/ https://paddleocr.bj.bcebos.com/dataset/icdar2015_benckmark.tar --no-check-certificate
+        cd ./train_data/ && tar xf icdar2015_benckmark.tar
+        ln -s ./icdar2015_benckmark ./icdar2015
+        cd ../
+    fi
+    if [[ ${model_name} =~ "ch_PP-OCRv2_det" ]];then
+        wget  -nc -P  ./pretrain_models/  https://paddleocr.bj.bcebos.com/dygraph_v2.0/ch/ch_ppocr_server_v2.0_det_train.tar --no-check-certificate
+        cd ./pretrain_models/ && tar xf ch_ppocr_server_v2.0_det_train.tar  && cd ../
+        rm -rf ./train_data/icdar2015
+        wget -nc -P ./train_data/ https://paddleocr.bj.bcebos.com/dataset/icdar2015_benckmark.tar --no-check-certificate
+        cd ./train_data/ && tar xf icdar2015_benckmark.tar
+        ln -s ./icdar2015_benckmark ./icdar2015
+        cd ../
     fi
     if [[ ${model_name} =~ "det_r50_vd_east_v2_0" ]]; then
         wget -nc -P ./pretrain_models/ https://paddleocr.bj.bcebos.com/dygraph_v2.0/en/det_r50_vd_east_v2.0_train.tar --no-check-certificate
         cd ./pretrain_models/ && tar xf det_r50_vd_east_v2.0_train.tar && cd ../
-        wget -nc -P ./train_data/ https://paddleocr.bj.bcebos.com/dygraph_v2.0/test/icdar2015.tar --no-check-certificate
-        cd ./train_data/ && tar xf icdar2015.tar && cd ../
+        rm -rf ./train_data/icdar2015
+        wget -nc -P ./train_data/ https://paddleocr.bj.bcebos.com/dataset/icdar2015_benckmark.tar --no-check-certificate
+        cd ./train_data/ && tar xf icdar2015_benckmark.tar
+        ln -s ./icdar2015_benckmark ./icdar2015
+        cd ../
     fi
-    if [[ ${model_name} =~ "det_r50_vd_pse_v2_0" ]];then
+    if [[ ${model_name} =~ "det_r50_db_v2.0" || ${model_name} =~ "det_r50_vd_pse_v2_0" ]];then
         wget -nc -P ./pretrain_models/ https://paddleocr.bj.bcebos.com/pretrained/ResNet50_vd_ssld_pretrained.pdparams --no-check-certificate
-        wget -nc -P ./train_data/ https://paddleocr.bj.bcebos.com/dygraph_v2.0/test/icdar2015.tar --no-check-certificate
-        cd ./train_data/ && tar xf icdar2015.tar && cd ../
+        rm -rf ./train_data/icdar2015
+        wget -nc -P ./train_data/ https://paddleocr.bj.bcebos.com/dataset/icdar2015_benckmark.tar --no-check-certificate
+        cd ./train_data/ && tar xf icdar2015_benckmark.tar
+        ln -s ./icdar2015_benckmark ./icdar2015
+        cd ../
     fi
     if [[ ${model_name} =~ "det_r18_db_v2_0" ]];then
         wget -nc -P ./pretrain_models/  https://paddleocr.bj.bcebos.com/pretrained/ResNet18_vd_pretrained.pdparams  --no-check-certificate
-        wget -nc -P ./train_data/ https://paddleocr.bj.bcebos.com/dygraph_v2.0/test/icdar2015.tar --no-check-certificate
-        cd ./train_data/ && tar xf icdar2015.tar && cd ../
+        rm -rf ./train_data/icdar2015
+        wget -nc -P ./train_data/ https://paddleocr.bj.bcebos.com/dataset/icdar2015_benckmark.tar --no-check-certificate
+        cd ./train_data/ && tar xf icdar2015_benckmark.tar
+        ln -s ./icdar2015_benckmark ./icdar2015
+        cd ../
+    fi
+    if [[ ${model_name} =~ "ch_ppocr_mobile_v2.0_rec" || ${model_name} =~ "ch_ppocr_server_v2.0_rec" || ${model_name} =~ "ch_PP-OCRv2_rec" || ${model_name} =~ "rec_mv3_none_bilstm_ctc_v2.0" || ${model_name} =~ "ch_PP-OCRv3_rec" ]];then
+        rm -rf ./train_data/ic15_data_benckmark
+        wget -nc -P ./train_data/ https://paddleocr.bj.bcebos.com/dataset/ic15_data_benckmark.tar --no-check-certificate
+        cd ./train_data/ && tar xf ic15_data_benckmark.tar
+        ln -s ./ic15_data_benckmark ./ic15_data
+        cd ../
+    fi
+    if [[ ${model_name} == "en_table_structure" ]];then
+        wget -nc -P ./pretrain_models/ https://paddleocr.bj.bcebos.com/dygraph_v2.1/table/en_ppocr_mobile_v2.0_table_structure_train.tar --no-check-certificate
+        cd ./pretrain_models/ && tar xf en_ppocr_mobile_v2.0_table_structure_train.tar  && cd ../
+        rm -rf ./train_data/pubtabnet
+        wget -nc -P ./train_data/ https://paddleocr.bj.bcebos.com/dataset/pubtabnet_benckmark.tar --no-check-certificate
+        cd ./train_data/ && tar xf pubtabnet_benckmark.tar
+        ln -s ./pubtabnet_benckmark ./pubtabnet
+        cd ../
+    fi
+    if [[ ${model_name} == "det_r50_dcn_fce_ctw_v2.0" ]]; then
+        wget -nc -P ./pretrain_models/ https://paddleocr.bj.bcebos.com/contribution/det_r50_dcn_fce_ctw_v2.0_train.tar --no-check-certificate
+        cd ./pretrain_models/ && tar xf det_r50_dcn_fce_ctw_v2.0_train.tar && cd ../
+        rm -rf ./train_data/icdar2015
+        wget -nc -P ./train_data/ https://paddleocr.bj.bcebos.com/dataset/icdar2015_benckmark.tar --no-check-certificate
+        cd ./train_data/ && tar xf icdar2015_benckmark.tar
+        ln -s ./icdar2015_benckmark ./icdar2015
+        cd ../
     fi
 fi
 
@@ -137,6 +189,10 @@ if [ ${MODE} = "lite_train_lite_infer" ];then
         wget -nc -P ./pretrain_models/ https://paddleocr.bj.bcebos.com/dygraph_v2.0/en/det_r50_vd_east_v2.0_train.tar --no-check-certificate
         cd ./pretrain_models/ && tar xf det_r50_vd_east_v2.0_train.tar && cd ../
     fi
+    if [ ${model_name} == "det_r50_dcn_fce_ctw_v2.0" ]; then
+        wget -nc -P ./pretrain_models/ https://paddleocr.bj.bcebos.com/contribution/det_r50_dcn_fce_ctw_v2.0_train.tar --no-check-certificate
+        cd ./pretrain_models/ && tar xf det_r50_dcn_fce_ctw_v2.0_train.tar & cd ../
+    fi
 
 elif [ ${MODE} = "whole_train_whole_infer" ];then
     wget -nc -P  ./pretrain_models/ https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/MobileNetV3_large_x0_5_pretrained.pdparams --no-check-certificate
@@ -363,6 +419,10 @@ elif [ ${MODE} = "whole_infer" ];then
         wget -nc -P ./inference/ https://paddleocr.bj.bcebos.com/dygraph_v2.0/en/det_r50_vd_east_v2.0_train.tar --no-check-certificate
         cd ./inference/ && tar xf det_r50_vd_east_v2.0_train.tar & cd ../
     fi
+    if [ ${model_name} == "det_r50_dcn_fce_ctw_v2.0" ]; then
+        wget -nc -P ./inference/ https://paddleocr.bj.bcebos.com/contribution/det_r50_dcn_fce_ctw_v2.0_train.tar --no-check-certificate
+        cd ./inference/ && tar xf det_r50_dcn_fce_ctw_v2.0_train.tar & cd ../
+    fi
     if [[ ${model_name} =~ "en_table_structure" ]];then
         wget -nc -P ./inference/ https://paddleocr.bj.bcebos.com/dygraph_v2.0/table/en_ppocr_mobile_v2.0_table_structure_infer.tar --no-check-certificate
         wget -nc -P ./inference/ https://paddleocr.bj.bcebos.com/dygraph_v2.0/table/en_ppocr_mobile_v2.0_table_det_infer.tar --no-check-certificate

test_tipc/readme.md

@@ -53,6 +53,8 @@
 | SRN    |rec_r50fpn_vd_none_srn         | 识别  | 支持 | 多机多卡 <br> 混合精度 | - | - |
 | NRTR   |rec_mtb_nrtr                   | 识别  | 支持 | 多机多卡 <br> 混合精度 | - | - |
 | SAR    |rec_r31_sar                    | 识别  | 支持 | 多机多卡 <br> 混合精度 | - | - |
+| SPIN   |rec_r32_gaspin_bilstm_att      | 识别  | 支持 | 多机多卡 <br> 混合精度 | - | - |
+| RobustScanner |rec_r31_robustscanner   | 识别  | 支持 | 多机多卡 <br> 混合精度 | - | - |
 | PGNet  |rec_r34_vd_none_none_ctc_v2.0  | 端到端| 支持 | 多机多卡 <br> 混合精度 | - | - |
 | TableMaster  |table_structure_tablemaster_train  | 表格识别| 支持 | 多机多卡 <br> 混合精度 | - | - |
 

tools/infer/predict_rec.py

@@ -68,15 +68,7 @@ class TextRecognizer(object):
                 'name': 'SARLabelDecode',
                 "character_dict_path": args.rec_char_dict_path,
                 "use_space_char": args.use_space_char
-            }
-        elif self.rec_algorithm == "RobustScanner":
-            postprocess_params = {
-                'name': 'SARLabelDecode',
-                "character_dict_path": args.rec_char_dict_path,
-                "use_space_char": args.use_space_char,
-                "rm_symbol": True
-            }
-                
+            }   
         elif self.rec_algorithm == 'ViTSTR':
             postprocess_params = {
                 'name': 'ViTSTRLabelDecode',
@@ -95,6 +87,13 @@ class TextRecognizer(object):
                 "character_dict_path": args.rec_char_dict_path,
                 "use_space_char": args.use_space_char
             }
+        elif self.rec_algorithm == "RobustScanner":
+            postprocess_params = {
+                'name': 'SARLabelDecode',
+                "character_dict_path": args.rec_char_dict_path,
+                "use_space_char": args.use_space_char,
+                "rm_symbol": True
+            }
         self.postprocess_op = build_post_process(postprocess_params)
         self.predictor, self.input_tensor, self.output_tensors, self.config = \
             utility.create_predictor(args, 'rec', logger)

tools/program.py

@@ -154,6 +154,24 @@ def check_xpu(use_xpu):
     except Exception as e:
         pass
 
+def to_float32(preds):
+    if isinstance(preds, dict):
+        for k in preds:
+            if isinstance(preds[k], dict) or isinstance(preds[k], list):
+                preds[k] = to_float32(preds[k])
+            else:
+                preds[k] = preds[k].astype(paddle.float32)
+    elif isinstance(preds, list):
+        for k in range(len(preds)):
+            if isinstance(preds[k], dict):
+                preds[k] = to_float32(preds[k])
+            elif isinstance(preds[k], list):
+                preds[k] = to_float32(preds[k])
+            else:
+                preds[k] = preds[k].astype(paddle.float32)
+    else:
+        preds = preds.astype(paddle.float32)
+    return preds
 
 def train(config,
           train_dataloader,
@@ -252,13 +270,19 @@ def train(config,
 
             # use amp
             if scaler:
-                with paddle.amp.auto_cast():
+                with paddle.amp.auto_cast(level='O2'):
                     if model_type == 'table' or extra_input:
                         preds = model(images, data=batch[1:])
                     elif model_type in ["kie", 'vqa']:
                         preds = model(batch)
                     else:
                         preds = model(images)
+                preds = to_float32(preds)
+                loss = loss_class(preds, batch)
+                avg_loss = loss['loss']
+                scaled_avg_loss = scaler.scale(avg_loss)
+                scaled_avg_loss.backward()
+                scaler.minimize(optimizer, scaled_avg_loss)
             else:
                 if model_type == 'table' or extra_input:
                     preds = model(images, data=batch[1:])
@@ -266,15 +290,8 @@ def train(config,
                     preds = model(batch)
                 else:
                     preds = model(images)
-
-            loss = loss_class(preds, batch)
-            avg_loss = loss['loss']
-
-            if scaler:
-                scaled_avg_loss = scaler.scale(avg_loss)
-                scaled_avg_loss.backward()
-                scaler.minimize(optimizer, scaled_avg_loss)
-            else:
+                loss = loss_class(preds, batch)
+                avg_loss = loss['loss']
                 avg_loss.backward()
                 optimizer.step()
             optimizer.clear_grad()

tools/train.py

@@ -157,6 +157,7 @@ def main(config, device, logger, vdl_writer):
         scaler = paddle.amp.GradScaler(
             init_loss_scaling=scale_loss,
             use_dynamic_loss_scaling=use_dynamic_loss_scaling)
+        model, optimizer = paddle.amp.decorate(models=model, optimizers=optimizer, level='O2', master_weight=True)
     else:
         scaler = None