Merge pull request #9 from qurator-spk/integrating_trocr_and_torch_ensembling_and_updating_characters_list-refactor

Integrating trocr and torch ensembling and updating characters list refactor

.gitignore

@@ -12,3 +12,5 @@ output.html
 *.sw?
 TAGS
 uv.lock
+/ignore
+*.log

src/eynollah/eynollah_ocr.py

@@ -343,7 +343,7 @@ class Eynollah_ocr(Eynollah):
         if out_image_with_text:
             image_text = Image.new("RGB", (img.shape[1], img.shape[0]), "white")
             draw = ImageDraw.Draw(image_text)
-            font = get_font()
+            font = get_font(font_size=40)
             
             for indexer_text, bb_ind in enumerate(total_bb_coordinates):
                 x_bb = bb_ind[0]

src/eynollah/training/cli.py

@@ -11,6 +11,7 @@ from .train import train_cli
 from .convert import convert_cli
 from .extract_line_gt import linegt_cli
 from .weights_ensembling import ensemble_cli
+from .generate_or_update_cnn_rnn_ocr_character_list import main as update_ocr_characters_cli
 
 @click.group('training')
 def main():
@@ -23,3 +24,4 @@ main.add_command(train_cli, 'train')
 main.add_command(convert_cli, 'convert')
 main.add_command(linegt_cli, 'export_textline_images_and_text')
 main.add_command(ensemble_cli, 'ensembling')
+main.add_command(update_ocr_characters_cli, 'generate_or_update_cnn_rnn_ocr_character_list')

src/eynollah/training/extract_line_gt.py

@@ -50,6 +50,12 @@ from ..utils import is_image_filename
     is_flag=True,
     help="if this parameter set to true, cropped textline images will not be masked with textline contour.",
 )
+@click.option(
+    "--exclude_vertical_lines",
+    "-exv",
+    is_flag=True,
+    help="if this parameter set to true, vertical textline images will be excluded.",
+)
 def linegt_cli(
     image,
     dir_in,
@@ -57,6 +63,7 @@ def linegt_cli(
     dir_out,
     pref_of_dataset,
     do_not_mask_with_textline_contour,
+    exclude_vertical_lines,
 ):
     assert bool(dir_in) ^ bool(image), "Set --dir-in or --image-filename, not both"
     if dir_in:
@@ -100,6 +107,9 @@ def linegt_cli(
 
                             x, y, w, h = cv2.boundingRect(textline_coords)
 
+                            if exclude_vertical_lines and h > 2 * w:
+                                continue
+
                             total_bb_coordinates.append([x, y, w, h])
 
                             img_poly_on_img = np.copy(img)

src/eynollah/training/generate_gt_for_training.py

@@ -6,6 +6,7 @@ from pathlib import Path
 from PIL import Image, ImageDraw, ImageFont
 import cv2
 import numpy as np
+from eynollah.utils.font import get_font
 
 from .gt_gen_utils import (
     filter_contours_area_of_image,
@@ -393,11 +394,15 @@ def visualize_reading_order(xml_file, dir_xml, dir_out, dir_imgs):
             layout = np.zeros( (y_len,x_len,3) ) 
             layout = cv2.fillPoly(layout, pts =co_text_all, color=(1,1,1))
                 
+            try:
                 img_file_name_with_format = find_format_of_given_filename_in_dir(dir_imgs, f_name)
                 img = cv2.imread(os.path.join(dir_imgs, img_file_name_with_format))
                 
                 overlayed = overlay_layout_on_image(layout, img, cx_ordered, cy_ordered, color, thickness)
                 cv2.imwrite(os.path.join(dir_out, f_name+'.png'), overlayed)
+            except:
+                pass
+
             
         else:
             img = np.zeros( (y_len,x_len,3) ) 
@@ -452,6 +457,7 @@ def visualize_textline_segmentation(xml_file, dir_xml, dir_out, dir_imgs):
         xml_file = os.path.join(dir_xml,ind_xml )
         f_name = Path(ind_xml).stem
         
+        try:
             img_file_name_with_format = find_format_of_given_filename_in_dir(dir_imgs, f_name)
             img = cv2.imread(os.path.join(dir_imgs, img_file_name_with_format))
                 
@@ -460,6 +466,8 @@ def visualize_textline_segmentation(xml_file, dir_xml, dir_out, dir_imgs):
             added_image = visualize_image_from_contours(co_tetxlines, img)
             
             cv2.imwrite(os.path.join(dir_out, f_name+'.png'), added_image)
+        except:
+            pass
 
         
         
@@ -509,15 +517,17 @@ def visualize_layout_segmentation(xml_file, dir_xml, dir_out, dir_imgs):
             f_name = Path(ind_xml).stem
         print(f_name, 'f_name')
         
+        try:
             img_file_name_with_format = find_format_of_given_filename_in_dir(dir_imgs, f_name)
             img = cv2.imread(os.path.join(dir_imgs, img_file_name_with_format))
                 
-        co_text, co_graphic, co_sep, co_img, co_table, co_map, co_noise, y_len, x_len = get_layout_contours_for_visualization(xml_file)
+            co_text, co_graphic, co_sep, co_img, co_table, co_map, co_music, co_noise, y_len, x_len = get_layout_contours_for_visualization(xml_file)
             
-        
+            added_image = visualize_image_from_contours_layout(co_text['paragraph'], co_text['header']+co_text['heading'], co_text['drop-capital'], co_sep, co_img, co_text['marginalia'], co_table, co_map, co_music, img)
-        added_image = visualize_image_from_contours_layout(co_text['paragraph'], co_text['header']+co_text['heading'], co_text['drop-capital'], co_sep, co_img, co_text['marginalia'], co_table, img)
 
             cv2.imwrite(os.path.join(dir_out, f_name+'.png'), added_image)
+        except:
+            pass
 
 
 
@@ -552,8 +562,8 @@ def visualize_ocr_text(xml_file, dir_xml, dir_out):
     else:
         xml_files_ind = [xml_file]
         
-    font_path = "Charis-7.000/Charis-Regular.ttf"  # Make sure this file exists!
+    ###font_path = "Charis-7.000/Charis-Regular.ttf"  # Make sure this file exists!
-    font = ImageFont.truetype(font_path, 40)
+    font = get_font(font_size=40)#ImageFont.truetype(font_path, 40)
         
     for ind_xml in tqdm(xml_files_ind):
         indexer = 0
@@ -590,11 +600,11 @@ def visualize_ocr_text(xml_file, dir_xml, dir_out):
                 
                 
                 is_vertical = h > 2*w  # Check orientation
-                font = fit_text_single_line(draw, ocr_texts[index], font_path, w, int(h*0.4) )
+                font = fit_text_single_line(draw, ocr_texts[index], w, int(h*0.4) )
                 
                 if is_vertical:
                     
-                    vertical_font = fit_text_single_line(draw, ocr_texts[index], font_path, h, int(w * 0.8))
+                    vertical_font = fit_text_single_line(draw, ocr_texts[index], h, int(w * 0.8))
 
                     text_img = Image.new("RGBA", (h, w), (255, 255, 255, 0))  # Note: dimensions are swapped
                     text_draw = ImageDraw.Draw(text_img)

src/eynollah/training/generate_or_update_cnn_rnn_ocr_character_list.py

@@ -0,0 +1,59 @@
+import os
+import numpy as np
+import json
+import click
+import logging
+
+
+
+def run_character_list_update(dir_labels, out, current_character_list):
+    ls_labels = os.listdir(dir_labels)
+    ls_labels = [ind for ind in ls_labels if ind.endswith('.txt')]
+    
+    if current_character_list:
+        with open(current_character_list, 'r') as f_name:
+            characters = json.load(f_name)
+            
+        characters = set(characters)
+    else:
+        characters = set()
+
+
+    for ind in ls_labels:
+        label  = open(os.path.join(dir_labels,ind),'r').read().split('\n')[0]
+            
+        for char in label:
+            characters.add(char)
+                
+                
+    characters = sorted(list(set(characters)))
+
+    with open(out, 'w') as f_name:
+        json.dump(characters, f_name)
+        
+
+@click.command()
+@click.option(
+    "--dir_labels",
+    "-dl",
+    help="directory of labels which are .txt files",
+    type=click.Path(exists=True, file_okay=False),
+    required=True,
+)
+@click.option(
+    "--current_character_list",
+    "-ccl",
+    help="existing character list in a .txt file that needs to be updated with a set of labels",
+    type=click.Path(exists=True, file_okay=True),
+    required=False,
+)
+@click.option(
+    "--out",
+    "-o",
+    help="An output .txt file where the generated or updated character list will be written",
+    type=click.Path(exists=False, file_okay=True),
+)
+
+def main(dir_labels, out, current_character_list):
+    run_character_list_update(dir_labels, out, current_character_list)
+    

src/eynollah/training/gt_gen_utils.py

@@ -8,7 +8,7 @@ from shapely import geometry
 from pathlib import Path
 from PIL import ImageFont
 from ocrd_utils import bbox_from_points
-
+from eynollah.utils.font import get_font
 
 KERNEL = np.ones((5, 5), np.uint8)
 NS = { 'pc': 'http://schema.primaresearch.org/PAGE/gts/pagecontent/2019-07-15'
@@ -18,7 +18,7 @@ with warnings.catch_warnings():
     warnings.simplefilter("ignore")
     
     
-def visualize_image_from_contours_layout(co_par, co_header, co_drop, co_sep, co_image, co_marginal, co_table, co_map, img):
+def visualize_image_from_contours_layout(co_par, co_header, co_drop, co_sep, co_image, co_marginal, co_table, co_map, co_music, img):
     alpha = 0.5
     
     blank_image = np.ones( (img.shape[:]), dtype=np.uint8) * 255
@@ -32,6 +32,7 @@ def visualize_image_from_contours_layout(co_par, co_header, co_drop, co_sep, co_
     col_marginal =  (106, 90, 205)
     col_table =  (0, 90, 205)
     col_map =  (90, 90, 205)
+    col_music =  (90, 90, 0)
     
     if len(co_image)>0:
         cv2.drawContours(blank_image, co_image, -1, col_image, thickness=cv2.FILLED)  # Fill the contour
@@ -60,6 +61,9 @@ def visualize_image_from_contours_layout(co_par, co_header, co_drop, co_sep, co_
     if len(co_map)>0:
         cv2.drawContours(blank_image, co_map, -1, col_map, thickness=cv2.FILLED)  # Fill the contour
         
+    if len(co_music)>0:
+        cv2.drawContours(blank_image, co_music, -1, col_music, thickness=cv2.FILLED)  # Fill the contour
+    
     img_final =cv2.cvtColor(blank_image, cv2.COLOR_BGR2RGB)
     
     added_image = cv2.addWeighted(img,alpha,img_final,1- alpha,0)
@@ -352,11 +356,11 @@ def get_textline_contours_and_ocr_text(xml_file):
                 ocr_textlines.append(ocr_text_in[0])
     return co_use_case, y_len, x_len, ocr_textlines
 
-def fit_text_single_line(draw, text, font_path, max_width, max_height):
+def fit_text_single_line(draw, text, max_width, max_height):
     initial_font_size = 50
     font_size = initial_font_size
     while font_size > 10:  # Minimum font size
-        font = ImageFont.truetype(font_path, font_size)
+        font = get_font(font_size=font_size)# ImageFont.truetype(font_path, font_size)
         text_bbox = draw.textbbox((0, 0), text, font=font)  # Get text bounding box
         text_width = text_bbox[2] - text_bbox[0]
         text_height = text_bbox[3] - text_bbox[1]
@@ -366,7 +370,7 @@ def fit_text_single_line(draw, text, font_path, max_width, max_height):
 
         font_size -= 2  # Reduce font size and retry
 
-    return ImageFont.truetype(font_path, 10)  # Smallest font fallback
+    return get_font(font_size=10)#ImageFont.truetype(font_path, 10)  # Smallest font fallback
 
 def get_layout_contours_for_visualization(xml_file):
     tree1 = ET.parse(xml_file, parser = ET.XMLParser(encoding='utf-8'))
@@ -389,6 +393,7 @@ def get_layout_contours_for_visualization(xml_file):
     co_img=[]
     co_table=[]
     co_map=[]
+    co_music=[]
     co_noise=[]
     
     types_text = []
@@ -631,6 +636,31 @@ def get_layout_contours_for_visualization(xml_file):
                         break
                 co_map.append(np.array(c_t_in))
                 
+        if tag.endswith('}MusicRegion') or tag.endswith('}musicregion'):
+            #print('sth')
+            for nn in root1.iter(tag):
+                c_t_in=[]
+                sumi=0
+                for vv in nn.iter():
+                    # check the format of coords
+                    if vv.tag==link+'Coords':
+                        coords=bool(vv.attrib)
+                        if coords:
+                            p_h=vv.attrib['points'].split(' ')
+                            c_t_in.append( np.array( [ [ int(x.split(',')[0]) , int(x.split(',')[1]) ]  for x in p_h] ) )
+                            break
+                        else:
+                            pass
+    
+    
+                    if vv.tag==link+'Point':
+                        c_t_in.append([ int(float(vv.attrib['x'])) , int(float(vv.attrib['y'])) ])
+                        sumi+=1
+                    #print(vv.tag,'in')
+                    elif vv.tag!=link+'Point' and sumi>=1:
+                        break
+                co_music.append(np.array(c_t_in))
+    
 
         if tag.endswith('}NoiseRegion') or tag.endswith('}noiseregion'):
             #print('sth')
@@ -656,7 +686,7 @@ def get_layout_contours_for_visualization(xml_file):
                     elif vv.tag!=link+'Point' and sumi>=1:
                         break
                 co_noise.append(np.array(c_t_in))
-    return co_text, co_graphic, co_sep, co_img, co_table, co_map, co_noise, y_len, x_len
+    return co_text, co_graphic, co_sep, co_img, co_table, co_map, co_music, co_noise, y_len, x_len
     
 def get_images_of_ground_truth(
         gt_list,
@@ -870,7 +900,7 @@ def get_images_of_ground_truth(
                 types_graphic_label = list(types_graphic_dict.values())
 
                 
-            labels_rgb_color = [ (0,0,0), (255,0,0), (255,125,0), (255,0,125), (125,255,125), (125,125,0), (0,125,255), (0,125,0), (125,125,125), (255,0,255), (125,0,125), (0,255,0),(0,0,255), (0,255,255), (255,125,125),  (0,125,125), (0,255,125), (255,125,255), (125,255,0), (125,255,255)]
+            labels_rgb_color = [ (0,0,0), (255,0,0), (255,125,0), (255,0,125), (125,255,125), (125,125,0), (0,125,255), (0,125,0), (125,125,125), (255,0,255), (125,0,125), (0,255,0),(0,0,255), (0,255,255), (255,125,125),  (0,125,125), (0,255,125), (255,125,255), (125,255,0), (125,255,255), (125,125,255)]
             
             
             region_tags=np.unique([x for x in alltags if x.endswith('Region')])   
@@ -882,6 +912,7 @@ def get_images_of_ground_truth(
             co_img=[]
             co_table=[]
             co_map=[]
+            co_music=[]
             co_noise=[]
             
             for tag in region_tags:
@@ -966,7 +997,7 @@ def get_images_of_ground_truth(
                                         if "rest_as_decoration" in types_graphic:
                                             types_graphic_without_decoration = [element for element in types_graphic if element!='rest_as_decoration' and element!='decoration']
                                             if len(types_graphic_without_decoration) == 0:
-                                                if "type" in nn.attrib:
+                                                #if "type" in nn.attrib:
                                                 c_t_in_graphic['decoration'].append( np.array( [ [ int(x.split(',')[0]) , int(x.split(',')[1]) ]  for x in p_h] ) )
                                             elif len(types_graphic_without_decoration) >= 1:
                                                 if "type" in nn.attrib:
@@ -974,12 +1005,14 @@ def get_images_of_ground_truth(
                                                         c_t_in_graphic[nn.attrib['type']].append( np.array( [ [ int(x.split(',')[0]) , int(x.split(',')[1]) ]  for x in p_h] ) )
                                                     else:
                                                         c_t_in_graphic['decoration'].append( np.array( [ [ int(x.split(',')[0]) , int(x.split(',')[1]) ]  for x in p_h] ) )
-                                                        
+                                                else:
+                                                    c_t_in_graphic['decoration'].append( np.array( [ [ int(x.split(',')[0]) , int(x.split(',')[1]) ]  for x in p_h] ) )
                                         else:
                                             if "type" in nn.attrib:
                                                 if nn.attrib['type'] in all_defined_graphic_types:
                                                     c_t_in_graphic[nn.attrib['type']].append( np.array( [ [ int(x.split(',')[0]) , int(x.split(',')[1]) ]  for x in p_h] ) ) 
                                             
+                                        
                                         break
                                     else:
                                         pass
@@ -989,7 +1022,7 @@ def get_images_of_ground_truth(
                                     if "rest_as_decoration" in types_graphic:
                                         types_graphic_without_decoration = [element for element in types_graphic if element!='rest_as_decoration' and element!='decoration']
                                         if len(types_graphic_without_decoration) == 0:
-                                            if "type" in nn.attrib:
+                                            #if "type" in nn.attrib:
                                             c_t_in_graphic['decoration'].append( [ int(float(vv.attrib['x'])) , int(float(vv.attrib['y'])) ] )
                                             sumi+=1
                                         elif len(types_graphic_without_decoration) >= 1:
@@ -1000,6 +1033,9 @@ def get_images_of_ground_truth(
                                                 else:
                                                     c_t_in_graphic['decoration'].append( [ int(float(vv.attrib['x'])) , int(float(vv.attrib['y'])) ] )
                                                     sumi+=1
+                                            else:
+                                                c_t_in_graphic['decoration'].append( [ int(float(vv.attrib['x'])) , int(float(vv.attrib['y'])) ] )
+                                                sumi+=1
                                                     
                                     else:
                                         if "type" in nn.attrib:
@@ -1119,6 +1155,32 @@ def get_images_of_ground_truth(
                                     break
                             co_map.append(np.array(c_t_in))
                             
+                if 'musicregion' in keys:
+                    if tag.endswith('}MusicRegion') or tag.endswith('}musicregion'):
+                        #print('sth')
+                        for nn in root1.iter(tag):
+                            c_t_in=[]
+                            sumi=0
+                            for vv in nn.iter():
+                                # check the format of coords
+                                if vv.tag==link+'Coords':
+                                    coords=bool(vv.attrib)
+                                    if coords:
+                                        p_h=vv.attrib['points'].split(' ')
+                                        c_t_in.append( np.array( [ [ int(x.split(',')[0]) , int(x.split(',')[1]) ]  for x in p_h] ) )
+                                        break
+                                    else:
+                                        pass
+                
+                
+                                if vv.tag==link+'Point':
+                                    c_t_in.append([ int(float(vv.attrib['x'])) , int(float(vv.attrib['y'])) ])
+                                    sumi+=1
+                                #print(vv.tag,'in')
+                                elif vv.tag!=link+'Point' and sumi>=1:
+                                    break
+                            co_music.append(np.array(c_t_in))
+            
                 if 'noiseregion' in keys:
                     if tag.endswith('}NoiseRegion') or tag.endswith('}noiseregion'):
                         #print('sth')
@@ -1195,6 +1257,10 @@ def get_images_of_ground_truth(
                     erosion_rate = 0#2
                     dilation_rate = 3#4
                     co_map, img_boundary = update_region_contours(co_map, img_boundary, erosion_rate, dilation_rate, y_len, x_len )
+                if "musicregion" in elements_with_artificial_class:
+                    erosion_rate = 0#2
+                    dilation_rate = 3#4
+                    co_music, img_boundary = update_region_contours(co_music, img_boundary, erosion_rate, dilation_rate, y_len, x_len )
                     
                     
                 
@@ -1222,6 +1288,8 @@ def get_images_of_ground_truth(
                     img_poly=cv2.fillPoly(img, pts =co_table, color=labels_rgb_color[ config_params['tableregion']])
                 if 'mapregion' in keys:  
                     img_poly=cv2.fillPoly(img, pts =co_map, color=labels_rgb_color[ config_params['mapregion']])
+                if 'musicregion' in keys:  
+                    img_poly=cv2.fillPoly(img, pts =co_music, color=labels_rgb_color[ config_params['musicregion']])
                 if 'noiseregion' in keys:  
                     img_poly=cv2.fillPoly(img, pts =co_noise, color=labels_rgb_color[ config_params['noiseregion']])
                     
@@ -1286,6 +1354,9 @@ def get_images_of_ground_truth(
                 if 'mapregion' in keys:
                     color_label = config_params['mapregion']
                     img_poly=cv2.fillPoly(img, pts =co_map, color=(color_label,color_label,color_label))
+                if 'musicregion' in keys:
+                    color_label = config_params['musicregion']
+                    img_poly=cv2.fillPoly(img, pts =co_music, color=(color_label,color_label,color_label))
                 if 'noiseregion' in keys:
                     color_label = config_params['noiseregion']
                     img_poly=cv2.fillPoly(img, pts =co_noise, color=(color_label,color_label,color_label))

src/eynollah/training/inference.py

@@ -132,15 +132,31 @@ class SBBPredict:
             self.model = Model(
                 self.model.get_layer(name = "image").input, 
                 self.model.get_layer(name = "dense2").output)
+            assert isinstance(self.model, Model)
+        elif self.task == "transformer-ocr":
+            import torch
+            from transformers import VisionEncoderDecoderModel, TrOCRProcessor
+
+            self.model = VisionEncoderDecoderModel.from_pretrained(self.model_dir)
+            self.processor = TrOCRProcessor.from_pretrained(self.model_dir)
+
+            if self.cpu:
+                self.device = torch.device('cpu')
+            else:
+                self.device = torch.device('cuda:0')
+
+            self.model.to(self.device)
+
+            assert isinstance(self.model, torch.nn.Module)
         else:
             self.model = load_model(self.model_dir, compile=False,
                                     custom_objects={"PatchEncoder": PatchEncoder,
                                                     "Patches": Patches})
+            assert isinstance(self.model, Model)
 
         ##if self.weights_dir!=None:
             ##self.model.load_weights(self.weights_dir)
             
-        assert isinstance(self.model, Model)
         if self.task != 'classification' and self.task != 'reading_order':
             last = self.model.layers[-1]
             self.img_height = last.output_shape[1]
@@ -231,6 +247,13 @@ class SBBPredict:
             pred_texts = pred_texts[0].replace("[UNK]", "")
             return pred_texts
 
+        elif self.task == "transformer-ocr":
+            from PIL import Image
+            image = Image.open(image_dir).convert("RGB")
+            pixel_values = self.processor(image, return_tensors="pt").pixel_values
+            generated_ids = self.model.generate(pixel_values.to(self.device))
+            return self.processor.batch_decode(generated_ids, skip_special_tokens=True)[0]
+            
             
         elif self.task == 'reading_order':
             img_height = self.config_params_model['input_height']
@@ -566,6 +589,8 @@ class SBBPredict:
                     cv2.imwrite(self.save,res)
             elif self.task == "cnn-rnn-ocr":
                 print(f"Detected text: {res}")
+            elif self.task == "transformer-ocr":
+                print(f"Detected text: {res}")
             else:
                 img_seg_overlayed, only_layout  = self.visualize_model_output(res, self.img_org, self.task)
                 if self.save:
@@ -672,7 +697,7 @@ def main(image, dir_in, model, patches, save, save_layout, ground_truth, xml_fil
     with open(os.path.join(model,'config.json')) as f:
         config_params_model = json.load(f)
     task = config_params_model['task']
-    if task not in ['classification', 'reading_order', "cnn-rnn-ocr"]:
+    if task not in ['classification', 'reading_order', "cnn-rnn-ocr", "transformer-ocr"]:
         assert not image or save, "For segmentation or binarization, an input single image -i also requires an output filename -s"
         assert not dir_in or out, "For segmentation or binarization, an input directory -di also requires an output directory -o"
     x = SBBPredict(image, dir_in, model, task, config_params_model,

src/eynollah/training/train.py

@@ -3,6 +3,7 @@ import sys
 import io
 import json
 import click
+from typing import Optional
 
 from tqdm import tqdm
 import requests
@@ -397,7 +398,7 @@ def run(_config,
         f1_threshold_classification=None,
         classification_classes_name=None,
         ## if task=cnn-rnn-ocr
-        characters_txt_file=None,
+        characters_txt_file: Optional[str]=None,
         color_padding_rotation=False,
         thetha_padd=None,
         bin_deg=False,
@@ -698,6 +699,79 @@ def run(_config,
             callbacks=callbacks,
             initial_epoch=index_start)
 
+    elif task=="transformer-ocr":
+        import torch
+        from torch.utils.data import Dataset as TorchDataset
+        from transformers import TrOCRProcessor, VisionEncoderDecoderModel, Seq2SeqTrainer, Seq2SeqTrainingArguments, default_data_collator
+        dir_img, dir_lab = get_dirs_or_files(dir_train)
+
+        if continue_training:
+            model = VisionEncoderDecoderModel.from_pretrained(dir_of_start_model)
+        else:
+            model = VisionEncoderDecoderModel.from_pretrained("microsoft/trocr-base-printed")
+            
+        processor = TrOCRProcessor.from_pretrained("microsoft/trocr-base-printed")
+        
+        
+        # Create a DataLoader
+        class TransformerOCRTorchDataset(TorchDataset):
+            """
+            Wraps preprocess_imgs in a format consumable by torch
+            """
+            def __init__(self, config, dir_img, dir_lab):
+                self.samples = preprocess_imgs(
+                        config,
+                        dir_img,
+                        dir_lab,
+                        processor=processor,
+                    )
+
+            def __len__(self):
+                return len(self.samples)
+
+            def __iter__(self):
+                yield from self.samples
+
+        dataset = TransformerOCRTorchDataset(_config, dir_img, dir_lab)
+        data_loader = torch.utils.data.DataLoader(dataset, batch_size=1)
+        train_dataset = data_loader.dataset
+            
+        # set special tokens used for creating the decoder_input_ids from the labels
+        model.config.decoder_start_token_id = processor.tokenizer.cls_token_id
+        model.config.pad_token_id = processor.tokenizer.pad_token_id
+        # make sure vocab size is set correctly
+        model.config.vocab_size = model.config.decoder.vocab_size
+
+        # set beam search parameters
+        model.config.eos_token_id = processor.tokenizer.sep_token_id
+        model.config.max_length = max_len
+        model.config.early_stopping = True
+        model.config.no_repeat_ngram_size = 3
+        model.config.length_penalty = 2.0
+        model.config.num_beams = 4
+        
+        
+        training_args = Seq2SeqTrainingArguments(
+            predict_with_generate=True,
+            num_train_epochs=n_epochs,
+            learning_rate=learning_rate,
+            per_device_train_batch_size=n_batch,
+            fp16=True, 
+            output_dir=dir_output,
+            logging_steps=2,
+            save_steps=save_interval,
+        )
+        
+        # instantiate trainer
+        trainer = Seq2SeqTrainer(
+            model=model,
+            tokenizer=processor.feature_extractor,
+            args=training_args,
+            train_dataset=train_dataset,
+            data_collator=default_data_collator,
+        )
+        trainer.train()
+
     elif task=='classification':
         if continue_training:
             model = load_model(dir_of_start_model, compile=False)
@@ -741,7 +815,7 @@ def run(_config,
             usable_checkpoints = [os.path.join(dir_output, 'model_{epoch:02d}'.format(epoch=epoch + 1))
                                   for epoch in usable_checkpoints]
             ens_path = os.path.join(dir_output, 'model_ens_avg')
-            run_ensembling(usable_checkpoints, ens_path)
+            run_ensembling(usable_checkpoints, ens_path, framework='tensorflow')
             _log.info("ensemble model saved under '%s'", ens_path)
 
     elif task=='reading_order':

src/eynollah/training/utils.py

@@ -789,7 +789,7 @@ def preprocess_imgs(config,
             lab = cv2.imread(os.path.join(dir_lab, img_name + '.png'))
         elif config['task'] == "enhancement":
             lab = cv2.imread(os.path.join(dir_lab, img))
-        elif config['task'] == "cnn-rnn-ocr":
+        elif config['task'] in ["cnn-rnn-ocr", "transformer-ocr"]:
             # assert lab == 'img_name + '.txt'
             with open(os.path.join(dir_lab, img_name + '.txt'), 'r') as f:
                 lab = f.read().split('\n')[0]
@@ -797,7 +797,7 @@ def preprocess_imgs(config,
             lab = None
 
         try:
-            if config['task'] == "cnn-rnn-ocr":
+            if config['task'] in ["cnn-rnn-ocr", "transformer-ocr"]:
                 yield from preprocess_img_ocr(img, img_name, lab, **config)
                 continue
             else:
@@ -1116,14 +1116,25 @@ def preprocess_img_ocr(
         number_of_backgrounds_per_image=None,
         list_all_possible_background_images=None,
         list_all_possible_foreground_rgbs=None,
+        task=None,
+        processor=None,
         **kwargs
 ):
     def scale_image(img):
         return scale_padd_image_for_ocr(img, input_height, input_width).astype(np.float32) / 255.
     #lab = vectorize_label(lab, char_to_num, padding_token, max_len)
     # now padded at Dataset.padded_batch
+    if task == 'cnn-rnn-ocr':
+        assert char_to_num, 'task is cnn-rnn-ocr, so preprocess_imgs_ocr should be passed "char_to_num"'
         lab = char_to_num(tf.strings.unicode_split(lab, input_encoding="UTF-8"))
-    yield scale_image(img), lab
+        yield_encoder = lambda x: x
+    elif task == 'transformer-ocr':
+        assert processor, 'task is transformer-ocr, so preprocess_imgs_ocr should be passed "processor"'
+        # TODO make max_length configurable again, if deemed sensible
+        lab = [l if l != self.processor.tokenizer.pad_token_id else -100
+                for l in processor.tokenizer(lab, padding="max_length", max_length=128).input_ids]
+        yield_encoder = lambda img_, lab_: {"pixel_values": processor(Image.fromarray(img_), return_tensors="pt").pixel_values.squeeze(), "labels": torch.tensor(lab_)}
+    yield yield_encoder(scale_image(img), lab)
     #to_yield = {"image": ret_x, "label": ret_y}
 
     if dir_img_bin:
@@ -1139,32 +1150,32 @@ def preprocess_img_ocr(
             for padd_col in padd_colors:
                 img_pad = do_padding_for_ocr(img, 1.2, padd_col)
                 img_rot = rotation_not_90_func_single_image(img_pad, thetha_ind)
-                yield scale_image(img_rot), lab
+                yield yield_encoder(scale_image(img_rot), lab)
     if rotation_not_90:
         for thetha_ind in thetha:
             img_rot = rotation_not_90_func_single_image(img, thetha_ind)
-            yield scale_image(img_rot), lab
+            yield yield_encoder(scale_image(img_rot), lab)
     if blur_aug:
         for blur_type in blur_k:
             img_blur = bluring(img, blur_type)
-            yield scale_image(img_blur), lab
+            yield yield_encoder(scale_image(img_blur), lab)
     if degrading:
         for deg_scale_ind in degrade_scales:
             img_deg = do_degrading(img, deg_scale_ind)
-            yield scale_image(img_deg), lab
+            yield yield_encoder(scale_image(img_deg), lab)
     if bin_deg:
         for deg_scale_ind in degrade_scales:
             img_deg  = do_degrading(img_bin_corr, deg_scale_ind)
-            yield scale_image(img_deg), lab
+            yield yield_encoder(scale_image(img_deg), lab)
     if brightening:
         for bright_scale_ind in brightness:
             img_bright  = do_brightening(img, bright_scale_ind)
-            yield scale_image(img_bright), lab
+            yield yield_encoder(scale_image(img_bright), lab)
     if padding_white:
         for padding_size in white_padds:
             for padd_col in padd_colors:
                 img_pad = do_padding_for_ocr(img, padding_size, padd_col)
-                yield scale_image(img_pad), lab
+                yield yield_encoder(scale_image(img_pad), lab)
     if adding_rgb_foreground:
         for i_n in range(number_of_backgrounds_per_image):
             background_image_chosen_name = random.choice(list_all_possible_background_images)
@@ -1178,7 +1189,7 @@ def preprocess_img_ocr(
             img_fg = \
                 return_binary_image_with_given_rgb_background_and_given_foreground_rgb(
                     img_bin_corr, img_rgb_background_chosen, foreground_rgb_chosen)
-            yield scale_image(img_fg), lab
+            yield yield_encoder(scale_image(img_fg), lab)
     if adding_rgb_background:
         for i_n in range(number_of_backgrounds_per_image):
             background_image_chosen_name = random.choice(list_all_possible_background_images)
@@ -1186,59 +1197,59 @@ def preprocess_img_ocr(
                 cv2.imread(dir_rgb_backgrounds + '/' + background_image_chosen_name)
             img_bg = \
                 return_binary_image_with_given_rgb_background(img_bin_corr, img_rgb_background_chosen)
-            yield scale_image(img_bg), lab
+            yield yield_encoder(scale_image(img_bg), lab)
     if binarization:
-        yield scale_image(img_bin_corr), lab
+        yield yield_encoder(scale_image(img_bin_corr), lab)
     if image_inversion:
         img_inv = invert_image(img_bin_corr)
-        yield scale_image(img_inv), lab
+        yield yield_encoder(scale_image(img_inv), lab)
     if channels_shuffling:
         for shuffle_index in shuffle_indexes:
             img_shuf = return_shuffled_channels(img, shuffle_index)
-            yield scale_image(img_shuf), lab
+            yield yield_encoder(scale_image(img_shuf), lab)
     if add_red_textlines:
         img_red = return_image_with_red_elements(img, img_bin_corr)
-        yield scale_image(img_red), lab
+        yield yield_encoder(scale_image(img_red), lab)
     if white_noise_strap:
         img_noisy = return_image_with_strapped_white_noises(img)
-        yield scale_image(img_noisy), lab
+        yield yield_encoder(scale_image(img_noisy), lab)
     if textline_skewing:
         for des_scale_ind in skewing_amplitudes:
             img_rot  = do_deskewing(img, des_scale_ind)
-            yield scale_image(img_rot), lab
+            yield yield_encoder(scale_image(img_rot), lab)
     if textline_skewing_bin:
         for des_scale_ind in skewing_amplitudes:
             img_rot  = do_deskewing(img_bin_corr, des_scale_ind)
-            yield scale_image(img_rot), lab
+            yield yield_encoder(scale_image(img_rot), lab)
     if textline_left_in_depth:
         img_warp  = do_direction_in_depth(img, 'left')
-        yield scale_image(img_warp), lab
+        yield yield_encoder(scale_image(img_warp), lab)
     if textline_left_in_depth_bin:
         img_warp  = do_direction_in_depth(img_bin_corr, 'left')
-        yield scale_image(img_warp), lab
+        yield yield_encoder(scale_image(img_warp), lab)
     if textline_right_in_depth:
         img_warp  = do_direction_in_depth(img, 'right')
-        yield scale_image(img_warp), lab
+        yield yield_encoder(scale_image(img_warp), lab)
     if textline_right_in_depth_bin:
         img_warp  = do_direction_in_depth(img_bin_corr, 'right')
-        yield scale_image(img_warp), lab
+        yield yield_encoder(scale_image(img_warp), lab)
     if textline_up_in_depth:
         img_warp  = do_direction_in_depth(img, 'up')
-        yield scale_image(img_warp), lab
+        yield yield_encoder(scale_image(img_warp), lab)
     if textline_up_in_depth_bin:
         img_warp  = do_direction_in_depth(img_bin_corr, 'up')
-        yield scale_image(img_warp), lab
+        yield yield_encoder(scale_image(img_warp), lab)
     if textline_down_in_depth:
         img_warp  = do_direction_in_depth(img, 'down')
-        yield scale_image(img_warp), lab
+        yield yield_encoder(scale_image(img_warp), lab)
     if textline_down_in_depth_bin:
         img_warp  = do_direction_in_depth(img_bin_corr, 'down')
-        yield scale_image(img_warp), lab
+        yield yield_encoder(scale_image(img_warp), lab)
     if pepper_aug:
         for pepper_ind in pepper_indexes:
             img_noisy = add_salt_and_pepper_noise(img, pepper_ind, pepper_ind)
-            yield scale_image(img_noisy), lab
+            yield yield_encoder(scale_image(img_noisy), lab)
     if pepper_bin_aug:
         for pepper_ind in pepper_indexes:
             img_noisy = add_salt_and_pepper_noise(img_bin_corr, pepper_ind, pepper_ind)
-            yield scale_image(img_noisy), lab
+            yield yield_encoder(scale_image(img_noisy), lab)

src/eynollah/training/weights_ensembling.py

@@ -1,4 +1,5 @@
 import os
+from typing import Optional
 from warnings import catch_warnings, simplefilter
 
 import click
@@ -11,33 +12,56 @@ from ocrd_utils import tf_disable_interactive_logs
 tf_disable_interactive_logs()
 import tensorflow as tf
 from tensorflow.keras.models import load_model
+import torch
+from transformers import VisionEncoderDecoderModel
 
 from ..patch_encoder import (
     PatchEncoder,
     Patches,
 )
     
-def run_ensembling(model_dirs, out_dir):
+def run_ensembling(dir_models, out, framework):
-    all_weights = []
+    ls_models = os.listdir(dir_models)
+    # model: Optional[VisionEncoderDecoderModel] = None
+    # model_name: Optional[str] = None
+    if framework=="torch":
+        models = []
+        sd_models = []
         
-    for model_dir in model_dirs:
+        for model_name in ls_models:
-        assert os.path.isdir(model_dir), model_dir
+            model = VisionEncoderDecoderModel.from_pretrained(os.path.join(dir_models, model_name))
-        model = load_model(model_dir, compile=False,
+            models.append(model)
-                           custom_objects=dict(PatchEncoder=PatchEncoder,
+            sd_models.append(model.state_dict())
-                                               Patches=Patches))
+        for key in sd_models[0]:
-        all_weights.append(model.get_weights())
+            sd_models[0][key] = sum(sd[key] for sd in sd_models) / len(sd_models)
             
-    new_weights = []
+        model.load_state_dict(sd_models[0])
-    for layer_weights in zip(*all_weights):
+        os.system("mkdir "+out)
-        layer_weights = np.array([np.array(weights).mean(axis=0)
+        torch.save(model.state_dict(), os.path.join(out, "pytorch_model.bin"))
-                                  for weights in zip(*layer_weights)])
+        os.system('cp ' + os.path.join(os.path.join(dir_models, model_name), "config.json") + " " + out)
-        new_weights.append(layer_weights)
+        
+    else:
+        weights=[]
+
+        for model_name in ls_models:
+            model =  load_model(os.path.join(dir_models, model_name), compile=False, custom_objects={'PatchEncoder':PatchEncoder, 'Patches': Patches})
+            weights.append(model.get_weights())
+            
+        new_weights = list()
+
+        for weights_list_tuple in zip(*weights):
+            new_weights.append(
+                [np.array(weights_).mean(axis=0)\
+                    for weights_ in zip(*weights_list_tuple)])
+            
+
+
+        new_weights = [np.array(x) for x in new_weights]
             
-    #model = tf.keras.models.clone_model(model)
         model.set_weights(new_weights)
-
+        model.save(out)
-    model.save(out_dir)
+        os.system('cp '+os.path.join(os.path.join(dir_models, model_name), "config.json") + " " + out)
-    os.system('cp ' + os.path.join(model_dirs[0], "config.json ") + out_dir + "/")
+        os.system('cp '+os.path.join(os.path.join(dir_models, model_name), "characters_org.txt") + " " + out)
 
 @click.command()
 @click.option(
@@ -56,12 +80,19 @@ def run_ensembling(model_dirs, out_dir):
     required=True,
     type=click.Path(exists=False, file_okay=False),
 )
-def ensemble_cli(in_, out):
+@click.option(
+    "--framework",
+    "-fw",
+    help="this parameter gets tensorflow or torch as model framework",
+    type=click.Choice(['torch', 'tensorflow']),
+    default="tensorflow"
+)
+
+def ensemble_cli(in_, out, framework):
     """
     mix multiple model weights
 
     Load a sequence of models and mix them into a single ensemble model
     by averaging their weights. Write the resulting model.
     """
-    run_ensembling(in_, out)
+    run_ensembling(in_, out, framework)
-

src/eynollah/utils/font.py

@@ -9,8 +9,8 @@ else:
     import importlib.resources as importlib_resources
 
 
-def get_font():
+def get_font(font_size):
     #font_path = "Charis-7.000/Charis-Regular.ttf"  # Make sure this file exists!
-    font = importlib_resources.files(__package__) / "../Charis-Regular.ttf"
+    font = importlib_resources.files(__package__) / "../Amiri-Regular.ttf"
     with importlib_resources.as_file(font) as font:
-        return ImageFont.truetype(font=font, size=40)
+        return ImageFont.truetype(font=font, size=font_size)

train/config_params_trocr.json

@@ -0,0 +1,82 @@
+{
+    "backbone_type" : "transformer",
+    "task": "transformer-ocr",
+    "n_classes" : 2,
+    "max_len": 192,
+    "n_epochs" : 1,
+    "input_height" : 32,
+    "input_width" : 512,
+    "weight_decay" : 1e-6,
+    "n_batch" : 1,
+    "learning_rate": 1e-5,
+    "save_interval": 1500,
+    "patches" : false,
+    "pretraining" : false,
+    "augmentation" : true,
+    "flip_aug" : false,
+    "blur_aug" : true,
+    "scaling" : false,
+    "adding_rgb_background": true,
+    "adding_rgb_foreground": true,
+    "add_red_textlines": true,
+    "white_noise_strap": true,
+    "textline_right_in_depth": true,
+    "textline_left_in_depth": true,
+    "textline_up_in_depth": true,
+    "textline_down_in_depth": true,
+    "textline_right_in_depth_bin": true,
+    "textline_left_in_depth_bin": true,
+    "textline_up_in_depth_bin": true,
+    "textline_down_in_depth_bin": true,
+    "bin_deg": true,
+    "textline_skewing": true,
+    "textline_skewing_bin": true,
+    "channels_shuffling": true,
+    "degrading": true,
+    "brightening": true,
+    "binarization" : true,
+    "pepper_aug": true,
+    "pepper_bin_aug": true,
+    "image_inversion": true,
+    "scaling_bluring" : false,
+    "scaling_binarization" : false,
+    "scaling_flip" : false,
+    "rotation": false,
+    "color_padding_rotation": true,
+    "padding_white": true,
+    "rotation_not_90": true,
+    "transformer_num_patches_xy": [56, 56],
+    "transformer_patchsize_x": 4,
+    "transformer_patchsize_y": 4,
+    "transformer_projection_dim": 64,
+    "transformer_mlp_head_units": [128, 64],
+    "transformer_layers": 1,
+    "transformer_num_heads": 1,
+    "transformer_cnn_first": false,
+    "blur_k" : ["blur","gauss","median"],
+    "padd_colors" : ["white", "black"],
+    "scales" : [0.6, 0.7, 0.8, 0.9],
+    "brightness" : [1.3, 1.5, 1.7, 2],
+    "degrade_scales" : [0.2, 0.4],
+    "pepper_indexes": [0.01, 0.005],
+    "skewing_amplitudes" : [5, 8],
+    "flip_index" : [0, 1, -1],
+    "shuffle_indexes" : [ [0,2,1], [1,2,0], [1,0,2] , [2,1,0]],
+    "thetha" : [0.1, 0.2, -0.1, -0.2],
+    "thetha_padd": [-0.6, -1, -1.4, -1.8, 0.6, 1, 1.4, 1.8],
+    "white_padds" : [0.1, 0.3, 0.5, 0.7, 0.9],
+    "number_of_backgrounds_per_image": 2,
+    "continue_training": false,
+    "index_start" : 0,
+    "dir_of_start_model" : " ", 
+    "weighted_loss": false,
+    "is_loss_soft_dice": false,
+    "data_is_provided": false,
+    "dir_train": "/home/vahid/extracted_lines/1919_bin/train",
+    "dir_eval": "/home/vahid/Documents/test/sbb_pixelwise_segmentation/test_label/pageextractor_test/eval_new",
+    "dir_output": "/home/vahid/extracted_lines/1919_bin/output",
+    "dir_rgb_backgrounds": "/home/vahid/Documents/1_2_test_eynollah/set_rgb_background",
+    "dir_rgb_foregrounds": "/home/vahid/Documents/1_2_test_eynollah/out_set_rgb_foreground",
+    "dir_img_bin": "/home/vahid/extracted_lines/1919_bin/images_bin"
+    
+}

train/requirements.txt

@@ -6,6 +6,10 @@ imutils
 scipy
 tensorflow-addons # for connected_components, depublished and only compatible with tensorflow < 2.16
 tensorflow < 2.16 # for tensorflow-addons, so only needed in training
-tf-keras # avoid keras 3 (also needs TF_USE_LEGACY_KERAS=1)
+tf-keras < 2.16 # avoid keras 3 (also needs TF_USE_LEGACY_KERAS=1)
-tf-data < 2.16 # for tensorflow-addons, so only needed in training
 protobuf < 5 # for tensorflow-addons, so only needed in training
+torch
+transformers <= 4.30.2 ; python_version < '3.10'
+transformers >= 5 ; python_version >= '3.10'
+eynollah-fork-tf2onnx == 1.17.0.post2
+ml_dtypes >= 0.5