Merge branch 'dygraph' of https://github.com/PaddlePaddle/PaddleOCR into update_requirements

ppocr/data/__init__.py

@@ -34,6 +34,7 @@ import paddle.distributed as dist
 from ppocr.data.imaug import transform, create_operators
 from ppocr.data.simple_dataset import SimpleDataSet
 from ppocr.data.lmdb_dataset import LMDBDataSet
+from ppocr.data.pgnet_dataset import PGDataSet
 
 __all__ = ['build_dataloader', 'transform', 'create_operators']
 
@@ -54,7 +55,7 @@ signal.signal(signal.SIGTERM, term_mp)
 def build_dataloader(config, mode, device, logger, seed=None):
     config = copy.deepcopy(config)
 
-    support_dict = ['SimpleDataSet', 'LMDBDataSet']
+    support_dict = ['SimpleDataSet', 'LMDBDataSet', 'PGDataSet']
     module_name = config[mode]['dataset']['name']
     assert module_name in support_dict, Exception(
         'DataSet only support {}'.format(support_dict))
@@ -72,14 +73,14 @@ def build_dataloader(config, mode, device, logger, seed=None):
     else:
         use_shared_memory = True
     if mode == "Train":
-        #Distribute data to multiple cards
+        # Distribute data to multiple cards
         batch_sampler = DistributedBatchSampler(
             dataset=dataset,
             batch_size=batch_size,
             shuffle=shuffle,
             drop_last=drop_last)
     else:
-        #Distribute data to single card
+        # Distribute data to single card
         batch_sampler = BatchSampler(
             dataset=dataset,
             batch_size=batch_size,

ppocr/data/imaug/__init__.py

@@ -28,6 +28,7 @@ from .label_ops import *
 
 from .east_process import *
 from .sast_process import *
+from .pg_process import *
 
 
 def transform(data, ops=None):

ppocr/data/imaug/label_ops.py

@@ -187,6 +187,32 @@ class CTCLabelEncode(BaseRecLabelEncode):
         return dict_character
 
 
+class E2ELabelEncode(BaseRecLabelEncode):
+    def __init__(self,
+                 max_text_length,
+                 character_dict_path=None,
+                 character_type='EN',
+                 use_space_char=False,
+                 **kwargs):
+        super(E2ELabelEncode,
+              self).__init__(max_text_length, character_dict_path,
+                             character_type, use_space_char)
+        self.pad_num = len(self.dict)  # the length to pad
+
+    def __call__(self, data):
+        texts = data['strs']
+        temp_texts = []
+        for text in texts:
+            text = text.lower()
+            text = self.encode(text)
+            if text is None:
+                return None
+            text = text + [self.pad_num] * (self.max_text_len - len(text))
+            temp_texts.append(text)
+        data['strs'] = np.array(temp_texts)
+        return data
+
+
 class AttnLabelEncode(BaseRecLabelEncode):
     """ Convert between text-label and text-index """
 
@@ -215,7 +241,7 @@ class AttnLabelEncode(BaseRecLabelEncode):
             return None
         data['length'] = np.array(len(text))
         text = [0] + text + [len(self.character) - 1] + [0] * (self.max_text_len
-                                                               - len(text) - 1)
+                                                               - len(text) - 2)
         data['label'] = np.array(text)
         return data
 
@@ -261,7 +287,7 @@ class SRNLabelEncode(BaseRecLabelEncode):
         if len(text) > self.max_text_len:
             return None
         data['length'] = np.array(len(text))
-        text = text + [char_num] * (self.max_text_len - len(text))
+        text = text + [char_num - 1] * (self.max_text_len - len(text))
         data['label'] = np.array(text)
         return data
 

ppocr/losses/__init__.py

@@ -29,10 +29,11 @@ def build_loss(config):
     # cls loss
     from .cls_loss import ClsLoss
 
+    # e2e loss
+    from .e2e_pg_loss import PGLoss
     support_dict = [
         'DBLoss', 'EASTLoss', 'SASTLoss', 'CTCLoss', 'ClsLoss', 'AttentionLoss',
-        'SRNLoss'
-    ]
+        'SRNLoss', 'PGLoss']
 
     config = copy.deepcopy(config)
     module_name = config.pop('name')